Interviews A Minecraft Movie: Piotr Karwas (VFX Supervisor) & Elizabeth “Liz” Bernard (Animation Supervisor) – Digital Domain By Vincent Frei - 20/05/2025 With 20+ years in film and advertising, Piotr Karwas has supervised VFX for top directors including David Fincher and Tim Burton. His recent work includes The Electric State, Elevation, he’s here today to talk about the Digital Domain work on A Minecraft Movie. After discussing The Electric State with us a few months ago, Elizabeth ‘Liz’ Bernard is back—this time to talk about her work on A Minecraft Movie. How did you and Digital Domain get involved on this show? Piotr Karwas: We were approached by the Warner Bros. VFX team as they were looking to bring on an additional vendor to help complete the remaining visual effects work. It was also an opportunity to be the best dad ever, since both of my kids are huge Minecraft fans. How was the collaboration with Director Jared Hess and VFX Supervisor Dan Lemmon? Piotr Karwas: I’ve known Dan Lemmon since our time working together at Digital Domain. We collaborated on several projects in the early 2000s, so it was a great opportunity to team up again. As for Jared, I’ve been a fan of his work ever since his first feature, Napoleon Dynamite. Years ago, I also had the chance to briefly work on his film Gentlemen Broncos. The blend of Jared’s signature humor with the global phenomenon of Minecraft was simply too exciting to pass up. I knew I wanted to be involved. Although Digital Domain joined the project later in the production process, we quickly got up to speed and delivered our work on time, despite some challenges. What are the sequences made by Digital Domain? Piotr Karwas: Our primary focus was the final battle sequence, but we also contributed to several fully live-action shots featuring set extensions and some CG animation, primarily around the entrance to the mine. Could you walk us through the process of creating the 500+ Piglins and Iron Golems clashing in that 360° battlefield? What were some of the major technical challenges in making that sequence come to life? Piotr Karwas: One of the key creative challenges was striking the right balance between honoring the iconic look and feel of the game while delivering a cinematic, action-packed experience. The goal was to blend a sense of realism through the movement and design of characters and environments with the charm and style of the Minecraft universe. Under the guidance of Jared and Dan, and in close collaboration with the other vendors, we were able to craft something truly unique. The battle unfolds during the Piglin invasion, as the sunny atmosphere of the Overworld begins to transform into a darker, more menacing environment reminiscent of the Nether. The entire battlefield was designed to function seamlessly in both visual states, capturing the warmth and cheer of the original Overworld, while also supporting its dramatic shift into a hellish landscape filled with fire and gloom. Additionally, the stylized world of Minecraft presented some technical hurdles, especially in large scale crowd sequences. The characters’ unconventional anatomy, limited range of motion, and signature fighting styles added an extra layer of complexity but also pushed us to be creative and innovative. Liz Bernard: We started out with 10 piglins, each of which had 2-6 costumes, plus 8 weapons and 2 torches. We were able to get a lot of variety out of that for our piglin horde. The golem army was smaller and more uniform, and golems are too powerful to be messing around with weapons. Instead, they prefer to use their iron arms to smash through their opponents, so we didn’t have as many different iterations, costumes, and such for the iron golems. In Animation, we moved as fast as possible to clean up and modify a big pile of motion capture so that we could provide our Crowds team with over 300 unique animation cycles and clips. Those clips included everything from golems doing windmill arm attacks, to piglins hacking at golem legs with axes and swords, to baby piglins running around in circles, freaking out. Although we started with motion capture for most of the piglin actions, transferring those actions from a full sized human mocap actor onto 10 different blocky piglins with giant angular heads and different body proportions meant each mocap clip needed a lot of adjustment. For example, piglins cannot raise their square little arms very high because their shoulders will quickly crash into the head. Once we had ploughed through that chunk of work, our Crowds team strung those actions together in Autodesk’s Golaem software to create the crowds seen in the backgrounds of our shots. In the mid-ground, we animated a custom set of longer battle vignettes that we could reuse in a number of shots by reorienting them to camera and offsetting their timing. And of course, there was quite a lot of hero animation to fill in the foreground and tell the story of the ebb and flow of battle. Minecraft is known for its blocky aesthetic. How was the animation rigging approached to stay true to that while enhancing the characters’ performance for live action? Liz Bernard: The piglins, hoglins, and wolves were rigged as pretty standard fleshy characters. The wolves had blocky chests and hips, which our Rigging team kept relatively stiff to preserve the original block shape even when we posed the characters in action. Because we were in a hurry on this fast-paced show, we needed to get rigs into Animation’s hands as quickly as possible, so we ended up relying on our shot modeling/sculpting team to adjust some of the characters’ bodies after animation was approved. They paid special attention to the final look of their shoulders, elbows, and knees, to make sure that the bend at each joint stayed as blocky as possible. The iron golems were a lot trickier to get right because they were made of stiff blocks of metal, had no deforming parts, and their joints were extremely simple. We discovered right away that making an iron golem walk meant instant intersections where the legs attach to the hips and the arms attach to the shoulders. So, to solve that problem and get the golems engaged in the battle, we had to give the animators complete control over every body part. This is not something we typically do because it’s easy to go overboard and end up “off model,” but in this scenario, it was absolutely required. This level of control gave the animators the option to offset body parts to achieve the necessary range of movement without visible intersections on screen or gaps in silhouette. Houdini was used to procedurally generate environments for the film. What were some of the most challenging aspects of creating such vast, dynamic landscapes? Piotr Karwas: The environments posed a unique challenge, not just because of their blocky nature, but because it was essential to make the Overworld feel like a tangible, real place. Some design choices were influenced by physical sets that had been built, while others drew inspiration from the real world. We focused on introducing a sense of natural imperfection, subtle elements like slightly uneven lines, worn edges, scuff marks, scattered debris, dirt, and even touches of realistic vegetation. All of this had to be carefully integrated into the strict, grid-based aesthetic of the Minecraft universe. Our environment team rose to the challenge and delivered a highly detailed immersive location, while working under a demanding production schedule. What were the biggest challenges in adapting Minecraft’s creatures and characters into a live-action setting while retaining their iconic charm? Liz Bernard: Because we came onto the project towards the end of shot production, we were able to use other vendors’ work as a rough guide for how our characters should move and behave. The Iron Golems in particular were a fun challenge, and although Sony had a few shots that we could refer to for the overall movement, it was up to us to establish their combat style. We researched how the golems fight in the game, which, as it turns out, is quite simple. We stuck with the game’s stiff legged walk, simple straight arm swings and skyward scoops to match the game’s iron golem attack style as closely as we could. However, purely matching the game wasn’t going to look quite right in a live action environment with human actors. So, on top of the simple performance, we layered weight shifts to counterbalance the heavy arms. We also added some additional step patterns to get them into position to attack piglins so that they would appear clever and more agile without straying too far from the game aesthetic. We also improved the contacts between the golems’ arms and each piglin to really sell the sense of effort and weight in each attack and hit. The destruction of the Overworld Portal seems like a monumental task. How was such an epic scene created, and what role did the environment and FX teams play in bringing it to life? Piotr Karwas: One of the biggest challenges was figuring out the mechanics of how large-scale destruction would play out in the Minecraft world. We had to carefully determine which elements would follow real-world physics and which could behave in a more stylized, game inspired way. We also spent considerable time designing the post-collapse remains to ensure they were visually compelling and cohesive for the staging of the shots that followed. The portal destruction sequence, in particular, appeared in the two widest shots of the entire sequence. These moments required both armies to be on screen, complete with fallen Iron Golems, defeated or pork-chopped Piglins, and digital doubles for every actor. It was a massive technical, logistical, and creative undertaking. We worked on these shots all the way to the final hour. Can you describe the bespoke zombification transformation pipeline developed for the Piglins? How did this process differ from traditional character transformations? Piotr Karwas: By the time we joined the project, the zombification effect had already been largely developed by Weta FX. One of our main challenges was matching their established look while ensuring visual consistency across both their shots and ours. At the same time, we knew we had to push the effect further. After all, this was the grand finale, where the evil army was being defeated. We also wanted the transformation to feel distinct, especially when applied to large numbers of Piglins. To achieve that, we went as far as building complete skeletal systems that could be revealed during the transformation. How was lighting handled to simulate firelit nighttime chaos? What tools or techniques were critical in creating that atmosphere? Piotr Karwas: To achieve the most realistic look, we used actual fire simulations to drive the lighting on both the characters and the environment. We initially started with just a few fires, but Jared was drawn to the striking visual accents and the warm hues they cast on the characters. As a result, we ended up populating the scenes with hundreds of fires. The air was also filled with atmospheric elements like smoke, ash, and embers, all contributing to a richly textured, apocalyptic feel in every shot. To manage this across numerous shots, we developed semi-automated methods to batch simulate the effects efficiently. How was humor incorporated into the animation, especially with subtle gags like pork chops dropping onto the battlefield or background antics? What approach was used to time these moments? Liz Bernard: We had a good time adding easter eggs and gags into this epic battle scene! Piotr let me know early on that Jared would welcome goofy touches, especially on background characters, so I made sure to pass that information on to my team of animators. What did I get back? Nose picks, butt scratches, belly slaps, and all sorts of goofy behaviour in the background. We usually didn’t center those things in the action unless it was called for in the script (in fact, Jared and Dan had to pull us back a few times), but if you look carefully, you’ll find maximum silliness all over the place. And, of course, I have to mention the Great Hog here. He was probably our funniest character and the most challenging to animate. His style of movement was dictated by his unusual proportions, in particular, his itty-bitty legs and long chunky arms. Those stumpy legs meant his stride was super short unless he broke out into a full quadruped gallop, which our team animated with extra oomph and effort to get a good laugh. The Great Hog’s low rider pants took a dive after one particularly strenuous sprint across the battlefield to attack Natalie. When he rose up on hind legs in anticipation of smashing her with his giant blocky fists, Jared wanted to make sure that the audience could see clearly the gag that his butt crack was showing. In VFX, we traditionally call a gradient of options that we would pitch to the client a “wedge.” In this case, we needed to get a read on how low to go with the pants before we simulated them in CFX, so we sent over a “wedge” for the pants and butt crack for Jared to pick from. In this case, I guess it was actually…a wedgie. How was the visual humor seamlessly integrated into the more intense action sequences, and was there a particular moment that captured this balance perfectly? Liz Bernard: The visual humour was baked into the action from the get-go, from game to storyboards to previsualization. I mean, there’s something inherently funny about dozens of piglins in shorts getting smacked high up into the air and then poof! turning into porkchops. When I welcomed each animator to the show, I made a point of mentioning that they were a little more free to tell jokes with pantomime on this show than they might be used to on more realistic productions. Because we were always in a mindset to make each shot as funny as possible without distracting from the story points, the flavour in animation dailies was always funny. My team pitched piglins getting run over by the Great Hog by accident, baby piglins riding on an iron golem attempting to attack its head, some utterly hilarious slow-mo action with exaggerated overlap a la Baywatch, a hoglin and piglin rider mugging for the camera as they gallop by in the foreground, wolves clamped on by their teeth to the waistband of the Great Hog, and a whole lot of over-acted melodramatic writhing as the piglins zombified after the portal collapsed. You don’t have to tell animators twice to be silly. With such a large-scale project, how did the animation and VFX teams collaborate to maintain a cohesive style across both departments? Were there any specific moments where this collaboration stood out? Liz Bernard: This was a ripsnorter of a project timeline, so all departments were hitting the ground running and developing their parts of the project concurrently. That’s always challenging and requires a lot of communication to pull off. We had a bullpen of experienced folks in our show leadership, and knowing what each other’s departments could and couldn’t do saved us a lot of time as we ramped up. For example, we were able to establish a single groundplane early in our animation schedule, which helped us avoid having to adjust the character’s feet contacts to an evolving environment even as artists were still building the overworld battlefield. The Environments team did a great job set dressing the environment after we had placed our animation and crowds’ characters, so that we did not need to worry about running through plants, pebbles, fires, or corpses. The FX department established a workflow with the Crowds team to “porkchop” any crowds piglin that got smacked by crowds’ iron golems. Setting these rules and criteria helped each department streamline their processes. And, our production team did an absolutely incredible job making sure that every department stayed on target, communicated their roadblocks and breakthroughs effectively downstream (and sometimes upstream!), and adjusted the schedule to compensate. I think the moment where this collaboration really shone was when we saw all of the zombification shots at the end of the sequence starting to fall into place. One day, it felt like we had all of them left to tackle, and the next, we had our methodology kinks worked out and they all started to drop like dominoes. Piotr Karwas: As Liz mentioned, the schedule was extremely tight, especially for executing the film’s grand finale. We had to move quickly and strategically, knowing that things would evolve and change throughout production. I’m incredibly grateful to our amazing team at Digital Domain. Across all four locations, Los Angeles, Vancouver, Montreal, and Hyderabad, our artists worked around the clock to bring it all to life. Yes, there were long hours, high stress, and the usual frustrations with technology that doesn’t always behave but in the end, it was all absolutely worth it. Seeing kids enjoy the movie so much made everything worthwhile. Chicken Jockey! Looking back on the project, what aspects of the visual effects are you most proud of? Piotr Karwas: Without a doubt, the Great Hog’s butt crack was the highlight for me. Liz Bernard: Just the sheer scale of this thing is wild. This is one of those “more is more” kinds of projects that don’t come around so often in our world of photorealistic visual effects. Nothing was supposed to be “real” here, everything was exaggerated, every frame we produced was chock full of crazy action, over the top humor, and easter eggs. I’m very proud of our animation team for embracing this hilarious show in a more cartoony style that we don’t often have the opportunity to tackle. And, man, it tickles me that the fans got such a huge kick out of it all when the movie hit theatres. What else can you ask for? How long have you worked on this show? Piotr Karwas: About six months Liz Bernard: Just six months from start to finish. What’s the VFX shots count? Piotr Karwas: Digital Domain delivered 187 shots not including omits. What is your next project? Piotr Karwas: A Minecraft Movie I hope. Jokes aside, I’m looking at a whole slate of different projects. Liz Bernard: All I can say right now is that it’s a science fiction film that comes out in 2026. Stay tuned! A big thanks for your time. WANT TO KNOW MORE?Digital Domain: Dedicated page about A Minecraft Movie on Digital Domain website. © Vincent Frei – The Art of VFX – 2025

May 20, 2025Ravie LakshmananCredential Theft / Browser Security An unknown threat actor has been attributed to creating several malicious Chrome Browser extensions since February 2024 that masquerade as seemingly benign utilities but incorporate covert functionality to exfiltrate data, receive commands, and execute arbitrary code. "The actor creates websites that masquerade as legitimate services, productivity tools, ad and media creation or analysis assistants, VPN services, crypto, banking and more to direct users to install corresponding malicious extensions on Google's Chrome Web Store (CWS)," the DomainTools Intelligence (DTI) team said in a report shared with The Hacker News. While the browser add-ons appear to offer the advertised features, they also enable credential and cookie theft, session hijacking, ad injection, malicious redirects, traffic manipulation, and phishing via DOM manipulation. Another factor that works in the extensions' favor is that they are configured to grant themselves excessive permissions via the manifest.json file, allowing them to interact with every site visited on the browser, execute arbitrary code retrieved from an attacker-controlled domain, perform malicious redirects, and even inject ads. The extensions have also been found to rely on the "onreset" event handler on a temporary document object model (DOM) element to execute code, likely in an attempt to bypass content security policy (CSP). Some of the identified lure websites impersonate legitimate products and services like DeepSeek, Manus, DeBank, FortiVPN, and Site Stats to entice users into downloading and installing the extensions. The add-ons then proceed to harvest browser cookies, fetch arbitrary scripts from a remote server, and set up a WebSocket connection to act as a network proxy for traffic routing. There is currently no visibility into how victims are redirected to the bogus sites, but DomainTools told the publication that it could involve usual methods like phishing and social media. "Because they appear in both Chrome Web Store and have adjacent websites, they can return from as results in normal web searches and for searches within the Chrome store," the company said. "Many of the lure websites used Facebook tracking IDs, which strongly suggests they are leveraging Facebook / Meta apps in some way to attract site visitors. Possibly through Facebook pages, groups, and even ads." As of writing, it's not known who is behind the campaign, although the threat actors have set up over 100 fake websites and malicious Chrome extensions. Google, for its part, has taken down the extensions. To mitigate risks, users are advised to stick with verified developers before downloading extensions, review requested permissions, scrutinize reviews, and refrain from using lookalike extensions. That said, it's also worth keeping in mind that ratings could be manipulated and artificially inflated by filtering negative user feedback. DomainTools, in an analysis published late last month, found evidence of extensions impersonating DeepSeek that redirected users providing low ratings (1-3 stars) to a private feedback form on the ai-chat-bot[.]pro domain, while sending those providing high ratings (4-5 stars) to the official Chrome Web Store review page. Found this article interesting? Follow us on Twitter  and LinkedIn to read more exclusive content we post. SHARE    

In the AI arms race, all the major players say they want to go nuclear.   Over the past year, the likes of Meta, Amazon, Microsoft, and Google have sent out a flurry of announcements related to nuclear energy. Some are about agreements to purchase power from existing plants, while others are about investments looking to boost unproven advanced technologies. This story is a part of MIT Technology Review’s series “Power Hungry: AI and our energy future,” on the energy demands and carbon costs of the artificial-intelligence revolution. These somewhat unlikely partnerships could be a win for both the nuclear power industry and large tech companies. Tech giants need guaranteed sources of energy, and many are looking for low-emissions ones to hit their climate goals. For nuclear plant operators and nuclear technology developers, the financial support of massive established customers could help keep old nuclear power plants open and push new technologies forward. “There [are] a lot of advantages to nuclear,” says Michael Terrell, senior director of clean energy and carbon reduction at Google. Among them, he says, are that it’s “clean, firm, carbon-free, and can be sited just about anywhere.” (Firm energy sources are those that provide constant power.)  But there’s one glaring potential roadblock: timing. “There are needs on different time scales,” says Patrick White, former research director at the Nuclear Innovation Alliance. Many of these tech companies will require large amounts of power in the next three to five years, White says, but building new nuclear plants can take close to a decade.  Some next-generation nuclear technologies, especially small modular reactors, could take less time to build, but the companies promising speed have yet to build their first reactors—and in some cases they are still years away from even modestly sized demonstrations.  This timing mismatch means that even as tech companies tout plans for nuclear power, they’ll actually be relying largely on fossil fuels, keeping coal plants open, and even building new natural gas plants that could stay open for decades. AI and nuclear could genuinely help each other grow, but the reality is that the growth could be much slower than headlines suggest.  AI’s need for speed The US alone has roughly 3,000 data centers, and current projections say the AI boom could add thousands more by the end of the decade. The rush could increase global data center power demand by as much as 165% by 2030, according to one recent analysis from Goldman Sachs. In the US, estimates from industry and academia suggest energy demand for data centers could be as high as 400 terawatt-hours by 2030—up from fewer than 100 terawatt-hours in 2020 and higher than the total electricity demand from the entire country of Mexico. There are indications that the data center boom might be decelerating, with some companies slowing or pausing some projects in recent weeks. But even the most measured projections, in analyses like one recent report from the International Energy Agency, predict that energy demand will increase. The only question is by how much.   Many of the same tech giants currently scrambling to build data centers have also set climate goals, vowing to reach net-zero emissions or carbon-free energy within the next couple of decades. So they have a vested interest in where that electricity comes from.  Nuclear power has emerged as a strong candidate for companies looking to power data centers while cutting emissions. Unlike wind turbines and solar arrays that generate electricity intermittently, nuclear power plants typically put out a constant supply of energy to the grid, which aligns well with what data centers need. “Data center companies pretty much want to run full out, 24/7,” says Rob Gramlich, president of Grid Strategies, a consultancy focused on electricity and transmission. It also doesn’t hurt that, while renewables are increasingly politicized and under attack by the current administration in the US, nuclear has broad support on both sides of the aisle.  The problem is how to build up nuclear capacity—existing facilities are limited, and new technologies will take time to build. In 2022, all the nuclear reactors in the US together provided around 800 terawatt-hours of electricity to the power grid, a number that’s been basically steady for the past two decades. To meet electricity demand from data centers expected in 2030 with nuclear power, we’d need to expand the fleet of reactors in the country by half. New nuclear news  Some of the most exciting headlines regarding the burgeoning relationship between AI and nuclear technology involve large, established companies jumping in to support innovations that could bring nuclear power into the 21st century.  In October 2024, Google signed a deal with Kairos Power, a next-generation nuclear company that recently received construction approval for two demonstration reactors from the US Nuclear Regulatory Commission (NRC). The company is working to build small, molten-salt-cooled reactors, which it says will be safer and more efficient than conventional technology. The Google deal is a long-term power-purchase agreement: The tech giant will buy up to 500 megawatts of electricity by 2035 from whatever plants Kairos manages to build, with the first one scheduled to come online by 2030.  Amazon is also getting involved with next-generation nuclear technology with a direct investment in Maryland-based X-energy. The startup is among those working to create smaller, more-standardized reactors that can be built more quickly and with less expense. In October, Amazon signed a deal with Energy Northwest, a utility in Washington state, that will see Amazon fund the initial phase of a planned X-energy small modular reactor project in the state. The tech giant will have a right to buy electricity from one of the modules in the first project, which could generate 320 megawatts of electricity and be expanded to generate as much as 960 megawatts. Many new AI-focused data centers under construction will require 500 megawatts of power or more, so this project might be just large enough to power a single site.  The project will help meet energy needs “beginning in the early 2030s,” according to Amazon’s website. X-energy is currently in the pre-application process with the NRC, which must grant approval before the Washington project can move forward. Solid, long-term plans could be a major help in getting next-generation technologies off the ground. “It’s going to be important in the next couple [of] years to see more firm commitments and actual money going out for these projects,” says Jessica Lovering, who cofounded the Good Energy Collective, a policy research organization that advocates for the use of nuclear energy.  However, these early projects won’t be enough to make a dent in demand. The next-generation reactors Amazon and Google are supporting are modestly sized demonstrations—the first commercial installations of new technologies. They won’t be close to the scale needed to meet the energy demand expected from new data centers by 2030.  To provide a significant fraction of the terawatt-hours of electricity large tech companies use each year, nuclear companies will likely need to build dozens of new plants, not just a couple of reactors.  Purchasing power  One approach to get around this mismatch is to target existing reactors.  Microsoft made headlines in this area last year when it signed a long-term power purchase agreement with Constellation, the owner of the Three Mile Island Unit 1 nuclear plant in Pennsylvania. Constellation plans to reopen one of the reactors at that site and rename it the Crane Clean Energy Center. The deal with Microsoft ensures that there will be a customer for the electricity from the plant, if it successfully comes back online. (It’s currently on track to do so in 2028.) “If you don’t want to wait a decade for new technology, one of the biggest tools that we have in our tool kit today is to support relicensing of operating power plants,” says Urvi Parekh, head of global energy for Meta. Older facilities can apply for 20-year extensions from the NRC, a process that customers buying the energy can help support as it tends to be expensive and lengthy, Parekh says.  While these existing reactors provide some opportunity for Big Tech to snap up nuclear energy now, a limited number are in good enough shape to extend or reopen.  In the US, 24 reactors have licenses that will be up for renewal before 2035, roughly a quarter of those in operation today. A handful of plants could potentially be reopened in addition to Three Mile Island, White says. Palisades Nuclear Plant in Michigan has received a $1.52 billion loan guarantee from the US Department of Energy to reopen, and the owner of the Duane Arnold Energy Center in Iowa has filed a request with regulators that could begin the reopening process. Some sites have reactors that could be upgraded to produce more power without building new infrastructure, adding a total of between two and eight gigawatts, according to a recent report from the Department of Energy. That could power a handful of moderately sized data centers, but power demand is growing for individual projects—OpenAI has suggested the need for data centers that would require at least five gigawatts of power.  Ultimately, new reactors will be needed to expand capacity significantly, whether they use established technology or next-generation designs. Experts tend to agree that neither would be able to happen at scale until at least the early 2030s.  In the meantime, decisions made today in response to this energy demand boom will have ripple effects for years. Most power plants can last for several decades or more, so what gets built today will likely stay on the grid through 2040 and beyond. Whether the AI boom will entrench nuclear energy, fossil fuels, or other sources of electricity on the grid will depend on what is introduced to meet demand now.  No individual technology, including nuclear power, is likely to be the one true solution. As Google’s Terrell puts it, everything from wind and solar, energy storage, geothermal, and yes, nuclear, will be needed to meet both energy demand and climate goals. “I think nuclear gets a lot of love,” he says. “But all of this is equally as important.”

Got a project that’s too bold to build? Submit your conceptual works, images and ideas for global recognition and print publication in the 2025 Vision Awards! The Main Entry deadline of June 6th is fast approach — submit your work today. To match or not to match? This dilemma has always been a burden for architects when designing extensions for listed buildings. Perhaps one of the most famous examples of a clashing extension and a listed building was Daniel Libeskind’s design for the Military History Museum in Dresden. Both praised and criticized by many, the modern extension aimed at spatially expressing Germany’s violent history. The architecture was characterized as “insensitive and inconsiderate” as well as “brave and bonkers.” Regardless of whether this specific architectural project was “the right move or not,” it poses a somewhat eternal architectural question: how should architects approach listed buildings, which oftentimes carry an immense amount of history and emotional resonance? The following seven projects reveal several tactics of dealing with delicate pieces of cultural heritage, while adding a new building extension to the mix. Aalt Stadhaus Differdange By witry & witry architecture urbanisme, Differdange, Luxembourg The existing building dates back from 1847, situated in between Church Notre-Dame-des Douleurs (which was torn down) and the local town hall. The project aim was to create a new urban landmark for Differdange. Long filigree columns were used along the west façade to create an interplay between old and new architectural elements that create a harmonic ensemble. In parallel, curated material choices such as tiles and wooden floors merge the two interiors crafting a coherent atmosphere. Void Practice Rooms By John McAslan + Partners,  London, United Kingdom Surrounded by three listed buildings, the Void acts as an extension for the Royal Academy of Music (RAM), hosting a general music practice room, an opera practice room and a new multifunctional RAM space. Preserving the overall aesthetic of the site, the Void Building’s structure and materiality draws from its historic context and internal functions. Specifically, the characteristic red brick and Baroque stone trimmings serve as an inspiration to the modern albeit contextually harmonious extension. Felix-Nussbaum-Haus Extension By Studio Libeskind, Osnabrück, Germany The Felix Nussbaum Haus was originally designed by Daniel Libeskind and was completed in 1998. The new extension, also designed by Libeskind in May 2011, provides an entrance hall to the museum, a new shop and a learning center. The design completely contradicts the existing structure and is comprised of a glass bridge that acts as a gateway. Through the specific colors and materials, the extension “blends” with both museums. The grey plaster provides a stark contrast to the Kunstgeschichtliche Museum and the Akzisehaus, while anthracite colored frames accentuate the series of openings. Additionally, the façade acts as a screen that carefully frames the geometries of the Museum openings, resulting in a grid that guides the overall composition. Marecollege By 24H-architecture, Leiden, The Netherlands Inspired by the school’s philosophical discourse, the concept for the new building extensions was derived from the anthroposophical philosophy, which claims that scientific theory does not describe reality, but the relations which belong to reality. The new building aimed at spatially recreating this theory through lemniscates. This intricate geometry becomes the heart of the design, accommodating many social spaces, auditoriums and a cafeteria. The new façade is as solid as its neighbor, made however out of grey masonry walls. Extension to the Historisches Museum By :mlzd, Bern, Switzerland The extension to the Historical Museum Bern, built by André Lambert in 1894, is composed by two distinct elements: a 1000m2 temporary exhibition hall located beneath a new civic square and a monolithic six story block along the southern side of the site that houses the Bern city archives, offices and a library. Architecturally, the structure interchanges between solid and transparent façades. Specifically, towards the square the building acts as a glazed modernist curtain wall that reveals all the activities that occur in its interior, while the south façade is made of solid, cast concrete punctured by small random openings. Rehabilitation and Extension of the Colani-UFO with shaft hall By SSP AG, Lünen, Germany In the technology centre of Lünen, over a former coal mine factory, hovers a  unique type of “UFO” building. Designed in 1985 by Luigi Colani the “UFO” has become a well-known landmark to the area, thus instigating a rejuvenation approach by the Academy of Applied Science center. The primary idea was to integrate the “UFO,” the shaft and the underlying building with a new extension, turning them into a multifunctional complex able to cater for a wide variety of venues. The design preserved that industrial charm of existing buildings, while maintaining a rather subtle form in comparison to the rest of the heterogenous buildings on site. Particularly the dark monolithic façade becomes a “quiet pause” in a somewhat complex and stimulating urban context. Museum De Fundatie By Bierman Henket architecten,  Zwolle, Netherlands Museum De Fundatie in Zwolle is situated on the border between the mediaeval city center and the open 19th century parkland and its canals. The new extension is a spectacular circular volume placed in the roof of the former Palace of Justice. The structure aesthetically contrasts the medieval façade of Blijmarkt, while however complimenting functionally the museum, acting as a new entrance. The extension – also called the Art Cloud – follows the substructure’s logic by being symmetrical in two directions, thus establishing a new identity for the complex’s urban presence. Its façade is clad with 55,000 three-dimensional ceramic elements, forming a “shimmering” surface that breaks down the original building’s solidity. Got a project that’s too bold to build? Submit your conceptual works, images and ideas for global recognition and print publication in the 2025 Vision Awards! The Main Entry deadline of June 6th is fast approach — submit your work today. Featured Image: Extension to the Historisches Museum by :mlzd, Bern, Switzerland The post Bold Meets Old: 7 Architectural Extensions that Clash with History appeared first on Journal.

More on this topicWhy is Lasdun’s towering homage to Mackintosh still under threat? The City of London gave the green light to the scheme last week (16 May) via delegated powers to kick-start redevelopment of Milton Gate, a 1980s block designed by National Theatre architect and former Royal Gold Medal winner Lasdun. Backed by developer Brookfield Properties, the part-demolition scheme will provide about 27,800m² of office space at 60 Chiswell Street, near the Barbican, through a demolition and façade replacement strategy which retains 70 per cent of the substructure. Historic England, which granted the Lasdun building a certificate of immunity from listing in 2023, had earlier raised concerns about the latest proposals, which, it argued in January, would ‘cause harm to several listed buildings’ in the Chiswell Street Conservation Area.Advertisement However, City of London planning officers approved the scheme, which they said would be compatible with the existing context, ‘being read as a well-layered piece of design which celebrates moments in the public realm’. Catherine Croft, director of heritage campaign group The Twentieth Century Society, described the decision to approve the scheme as ‘hugely regrettable’ in comments to the AJ. Croft argued: ‘The consented scheme will totally destroy the architecturally and culturally significant elements of a fine building. The City will lose a major work by an outstanding architect and further erode its distinctive character, despite the building being well over the 30-year threshold for listing at Grade II.’ Croft continued: ‘Milton Gate is a clear example of a building that should have been saved by listing; C20 Society put it forward and engaged with the developer on several occasions, championing its architectural significance.’ ‘The proposed removal of the notable blue-glazed exterior form of 60 Chiswell Street would also remove all its architectural and historic interest, while the maximisation of the site through extensions would intrude upon and curtail open views from the south courtyard and along Chiswell Street,’ said the heritage body.Advertisement Brookfield said the scheme would appeal to ‘modern occupiers looking for large, efficient floorplates and outstanding environmental, social and governance credentials’, while reducing its environmental impact. The all-electric scheme is aiming for net zero in operation and aims to achieve EPC A, BREEAM Outstanding and NABERS 5* certificates. AHMM’s proposals include repositioning an entrance on the corner of Moor Lane and Ropemaker Street; a statement double-height reception; and a pocket park. The ground floor will also accommodate food and beverage uses. Simon Allford, founding director of Allford Hall Monaghan Morris, said the scheme was ‘the result of Brookfield Properties and AHMM’s shared ambition to create an innovative, low-carbon reinvention of an existing building’. He added: ‘The activated ground floor engages with the historic streetscape, whilst the generous office volumes above open onto, cascading balconies and verdant terraces, delivering a new green edge to the city’s historic skyline.’ A timeline for completion is unknown.

A shared language for product teams to build usable, intelligent and safe GenAI experiences beyond just the modelGenerative AI introduces a new way for humans to interact with systems by focusing on intent-based outcome specification. GenAI introduces novel challenges because its outputs are probabilistic, requires understanding of variability, memory, errors, hallucinations and malicious use which brings an essential need to build principles and design patterns as described by IBM.Moreover, any AI product is a layered system where LLM is just one ingredient and memory, orchestration, tool extensions, UX and agentic user-flows builds the real magic!This article is my research and documentation of evolving GenAI design patterns that provide a shared language for product managers, data scientists, and interaction designers to create products that are human-centred, trustworthy and safe. By applying these patterns, we can bridge the gap between user needs, technical capabilities and product development process.Here are 21 GenAI UX patternsGenAI or no GenAIConvert user needs to data needsAugment or automateDefine level of automationProgressive AI adoptionLeverage mental modelsConvey product limitsDisplay chain of thought (CoT)Leverage multiple outputsProvide data sourcesConvey model confidenceDesign for memory and recallProvide contextual input parametersDesign for coPilot, co-Editing or partial automationDefine user controls for AutomationDesign for user input error statesDesign for AI system error statesDesign to capture user feedbackDesign for model evaluationDesign for AI safety guardrailsCommunicate data privacy and controls1. GenAI or no GenAIEvaluate whether GenAI improves UX or introduces complexity. Often, heuristic-based (IF/Else) solutions are easier to build and maintain.Scenarios when GenAI is beneficialTasks that are open-ended, creative and augments user.E.g., writing prompts, summarizing notes, drafting replies.Creating or transforming complex outputs (e.g., images, video, code).E.g., converting a sketch into website code.Where structured UX fails to capture user intent.Scenarios when GenAI should be avoidedOutcomes that must be precise, auditable or deterministic. E.g., Tax forms or legal contracts.Users expect clear and consistent information.E.g. Open source software documentationHow to use this patternDetermine the friction points in the customer journeyAssess technology feasibility: Determine if AI can address the friction point. Evaluate scale, dataset availability, error risk assessment and economic ROI.Validate user expectations: - Determine if the AI solution erodes user expectations by evaluating whether the system augments human effort or replaces it entirely, as outlined in pattern 3, Augment vs. automate. - Determine if AI solution erodes pattern 6, Mental models2. Convert user needs to data needsThis pattern ensures GenAI development begins with user intent and data model required to achieve that. GenAI systems are only as good as the data they’re trained on. But real users don’t speak in rows and columns, they express goals, frustrations, and behaviours. If teams fail to translate user needs into structured, model-ready inputs, the resulting system or product may optimise for the wrong outcomes and thus user churn.How to use this patternCollaborate as a cross-functional team of PMs, Product designers and Data Scientists and align on user problems worth solving.Define user needs by using triangulated research: Qualitative (Market Reports, Surveys or Questionnaires) + Quantitative (User Interviews, Observational studies) + Emergent (Product reviews, Social listening etc.) and synthesising user insights using JTBD framework, Empathy Map to visualise user emotions and perspectives. Value Proposition Canvas to align user gains and pains with featuresDefine data needs and documentation by selecting a suitable data model, perform gap analysis and iteratively refine data model as needed. Once you understand the why, translate it into the what for the model. What features, labels, examples, and contexts will your AI model need to learn this behaviour? Use structured collaboration to figure out.3. Augment vs automateOne of the critical decisions in GenAI apps is whether to fully automate a task or to augment human capability. Use this pattern to to align with user intent and control preferences with the technology.Automation is best for tasks users prefer to delegate especially when they are tedious, time-consuming or unsafe. E.g., Intercom FinAI automatically summarizes long email threads into internal notes, saving time on repetitive, low-value tasks.Augmentation enhances tasks users want to remain involved in by increasing efficiency, increase creativity and control. E.g., Magenta Studio in Abelton support creative controls to manipulate and create new music.How to use this patternTo select the best approach, evaluate user needs and expectations using research synthesis tools like empathy map (visualise user emotions and perspectives) and value proposition canvas (to understand user gains and pains)Test and validate if the approach erodes user experience or enhances it.4. Define level of automationIn AI systems, automation refers to how much control is delegated to the AI vs user. This is a strategic UX pattern to decide degree of automation based upon user pain-point, context scenarios and expectation from the product.Levels of automationNo automation (AI assists but user decides)The AI system provides assistance and suggestions to the user but requires the user to make all the decisions. E.g., Grammarly highlights grammar issues but the user accepts or rejects corrections.Partial automation/ co-pilot/ co-editor (AI acts with user oversight)The AI initiates actions or generates content, but the user reviews or intervenes as needed. E.g., GitHub Copilot suggest code that developers can accept, modify, or ignore.Full automation (AI acts independently)The AI system performs tasks without user intervention, often based on predefined rules, tools and triggers. Full automation in GenAI are often referred to as Agentic systems. E.g., Ema can autonomously plan and execute multi-step tasks like researching competitors, generating a report and emailing it without user prompts or intervention at each step.How to use this patternEvaluate user pain point to be automated and risk involved: Automating tasks is most effective when the associated risk is low without severe consequences in case of failure. Low-risk tasks such as sending automated reminders, promotional emails, filtering spam emails or processing routine customer queries can be automated with minimal downside while saving time and resources. High-risk tasks such as making medical diagnoses, sending business-critical emails, or executing financial trades requires careful oversight due to the potential for significant harm if errors occur.Evaluate and design for particular automation level: Evaluate if user pain point should fall under — No Automation, Partial Automation or Full Automation based upon user expectations and goals.Define user controls for automation (refer pattern 15)5. Progressive GenAI adoptionWhen users first encounter a product built on new technology, they often wonder what the system can and can’t do, how it works and how they should interact with it.This pattern offers multi-dimensional strategy to help user onboard an AI product or feature, mitigate errors, aligns with user readiness to deliver an informed and human-centered UX.How to use this patternThis pattern is a culmination of many other patternsFocus on communicating benefits from the start: Avoid diving into details about the technology and highlight how the AI brings new value.Simplify the onboarding experience Let users experience the system’s value before asking data-sharing preferences, give instant access to basic AI features first. Encourage users to sign up later to unlock advanced AI features or share more details. E.g., Adobe FireFly progressively onboards user with basic to advance AI featuresDefine level of automation (refer pattern 4) and gradually increase autonomy or complexity.Provide explainability and trust by designing for errors (refer pattern 16 and 17).Communicate data privacy and controls (refer pattern 21) to clearly convey how user data is collected, stored, processed and protected.6. Leverage mental modelsMental models help user predict how a system (web, application or other kind of product) will work and, therefore, influence how they interact with an interface. When a product aligns with a user’s existing mental models, it feels intuitive and easy to adopt. When it clashes, it can cause frustration, confusion, or abandonment​.E.g. Github Copilot builds upon developers’ mental models from traditional code autocomplete, easing the transition to AI-powered code suggestionsE.g. Adobe Photoshop builds upon the familiar approach of extending an image using rectangular controls by integrating its Generative Fill feature, which intelligently fills the newly created space.How to use this patternIdentifying and build upon existing mental models by questioningWhat is the user journey and what is user trying to do?What mental models might already be in place?Does this product break any intuitive patterns of cause and effect?Are you breaking an existing mental model? If yes, clearly explain how and why. Good onboarding, microcopy, and visual cues can help bridge the gap.7. Convey product limitsThis pattern involves clearly conveying what an AI model can and cannot do, including its knowledge boundaries, capabilities and limitations.It is helpful to builds user trust, sets appropriate expectations, prevents misuse, and reduces frustration when the model fails or behaves unexpectedly.How to use this patternExplicitly state model limitations: Show contextual cues for outdated knowledge or lack of real-time data. E.g., Claude states its knowledge cutoff when the question falls outside its knowledge domainProvide fallbacks or escalation options when the model cannot provide a suitable output. E.g., Amazon Rufus when asked about something unrelated to shopping, says “it doesn’t have access to factual information and, I can only assists with shopping related questions and requests”Make limitations visible in product marketing, onboarding, tooltips or response disclaimers.8. Display chain of thought (CoT)In AI systems, chain-of-thought (CoT) prompting technique enhances the model’s ability to solve complex problems by mimicking a more structured, step-by-step thought process like that of a human.CoT display is a UX pattern that improves transparency by revealing how the AI arrived at its conclusions. This fosters user trust, supports interpretability, and opens up space for user feedback especially in high-stakes or ambiguous scenarios.E.g., Perplexity enhances transparency by displaying its processing steps helping users understand the thoughtful process behind the answers.E.g., Khanmigo an AI Tutoring system guides students step-by-step through problems, mimicking human reasoning to enhance understanding and learning.How to use this patternShow status like “researching” and “reasoning to communicate progress, reduce user uncertainty and wait times feel shorter.Use progressive disclosure: Start with a high-level summary, and allow users to expand details as needed.Provide AI tooling transparency: Clearly display external tools and data sources the AI uses to generate recommendations.Show confidence & uncertainty: Indicate AI confidence levels and highlight uncertainties when relevant.9. Leverage multiple outputsGenAI can produce varied responses to the same input due to its probabilistic nature. This pattern exploits variability by presenting multiple outputs side by side. Showing diverse options helps users creatively explore, compare, refine or make better decisions that best aligns with their intent. E.g., Google Gemini provides multiple options to help user explore, refine and make better decisions.How to use this patternExplain the purpose of variation: Help users understand that differences across outputs are intentional and meant to offer choice.Enable edits: Let users rate, select, remix, or edit outputs seamlessly to shape outcomes and provide feedback. E.g., Midjourney helps user adjust prompt and guide your variations and edits using remix10. Provide data sourcesArticulating data sources in a GenAI application is essential for transparency, credibility and user trust. Clearly indicating where the AI derives its knowledge helps users assess the reliability of responses and avoid misinformation.This is especially important in high stakes factual domains like healthcare, finance or legal guidance where decisions must be based on verified data.How to use this patternCite credible sources inline: Display sources as footnotes, tooltips, or collapsible links. E.g., NoteBookLM adds citations to its answers and links each answer directly to the part of user’s uploaded documents.Disclose training data scope clearly: For generative tools (text, images, code), offer a simple explanation of what data the model was trained on and what wasn’t included. E.g., Adobe Firefly discloses that its Generative Fill feature is trained on stock imagery, openly licensed work and public domain content where the copyright has expired.Provide source-level confidence:In cases where multiple sources contribute, visually differentiate higher-confidence or more authoritative sources.11. Convey model confidenceAI-generated outputs are probabilistic and can vary in accuracy. Showing confidence scores communicates how certain the model is about its output. This helps users assess reliability and make better-informed decisions.How to use this patternAssess context and decision stakes: Showing model confidence depends on the context and its impact on user decision-making. In high-stakes scenarios like healthcare, finance or legal advice, displaying confidence scores are crucial. However, in low stake scenarios like AI-generated art or storytelling confidence may not add much value and could even introduce unnecessary confusion.Choose the right visualization: If design research shows that displaying model confidence aids decision-making, the next step is to select the right visualization method. Percentages, progress bars or verbal qualifiers (“likely,” “uncertain”) can communicate confidence effectively. The apt visualisation method depends on the application’s use-case and user familiarity. E.g., Grammarly uses verbal qualifiers like “likely” to the content it generated along with the userGuide user action during low confidence scenarios: Offer paths forward such as asking clarifying questions or offering alternative options.12. Design for memory and recallMemory and recall is an important concept and design pattern that enables the AI product to store and reuse information from past interactions such as user preferences, feedback, goals or task history to improve continuity and context awareness.Enhances personalization by remembering past choices or preferencesReduces user burden by avoiding repeated input requests especially in multi-step or long-form tasksSupports complex tasks like longitudinal workflows like in project planning, learning journeys by referencing or building on past progress.Memory used to access information can be ephemeral (short-term within a session) or persistent (long-term across sessions) and may include conversational context, behavioural signals, or explicit inputs.How to use this patternDefine the user context and choose memory typeChoose memory type like ephemeral or persistent or both based upon use case. A shopping assistant might track interactions in real time without needing to persist data for future sessions whereas personal assistants need long-term memory for personalization.Use memory intelligently in user interactionsBuild base prompts for LLM to recall and communicate information contextually (E.g., “Last time you preferred a lighter tone. Should I continue with that?”).Communicate transparency and provide controlsClearly communicate what’s being saved and let users view, edit or delete stored memory. Make “delete memories” an accessible action. E.g. ChatGPT offers extensive controls across it’s platform to view, update, or delete memories anytime.13. Provide contextual input parametersContextual Input parameters enhance the user experience by streamlining user interactions and gets to user goal faster. By leveraging user-specific data, user preferences or past interactions or even data from other users who have similar preferences, GenAI system can tailor inputs and functionalities to better meet user intent and decision making.How to use this patternLeverage prior interactions: Pre-fill inputs based on what the user has previously entered. Refer pattern 12, Memory and recall.Use auto complete or smart defaults: As users type, offer intelligent, real-time suggestions derived from personal and global usage patterns. E.g., Perplexity offers smart next query suggestions based on your current query thread.Suggest interactive UI widgets: Based upon system prediction, provide tailored input widgets like toasts, sliders, checkboxes to enhance user input. E.g., ElevenLabs allows users to fine-tune voice generation settings by surfacing presets or defaults.14. Design for co-pilot / co-editing / partial automationCo-pilot is an augmentation pattern where AI acts as a collaborative assistant, offering contextual and data-driven insights while the user remains in control. This design pattern is essential in domains like strategy, ideating, writing, designing or coding where outcomes are subjective, users have unique preferences or creative input from the user is critical.Co-pilot speed up workflows, enhance creativity and reduce cognitive load but the human retains authorship and final decision-making.How to use this patternEmbed inline assistance: Place AI suggestions contextually so users can easily accept, reject or modify them. E.g., Notion AI helps you draft, summarise and edit content while you control the final version.Save user intent and creative direction: Let users guide the AI with input like goals, tone, or examples, maintaining authorship and creative direction. E.g., Jasper AI allows users to set brand voice and tone guidelines, helping structure AI output to better match the user’s intent.15. Design user controls for automationBuild UI-level mechanisms that let users manage or override automation based upon user goals, context scenarios or system failure states.No system can anticipate all user contexts. Controls give users agency and keep trust intact even when the AI gets it wrong.How to use this patternUse progressive disclosure: Start with minimal automation and allow users to opt into more complex or autonomous features over time. E.g., Canva Magic Studio starts with simple AI suggestions like text or image generation then gradually reveals advanced tools like Magic Write, AI video scenes and brand voice customisation.Give users automation controls: UI controls like toggles, sliders, or rule-based settings to let users choose when and how automation can be controlled. E.g., Gmail lets users disable Smart Compose.Design for automation error recovery: Give users correction when AI fails (false positives/negatives). Add manual override, undo, or escalate options to human support. E.g., GitHub Copilot suggests code inline, but developers can easily reject, modify or undo suggestions when output is off.16. Design for user input error statesGenAI systems often rely on interpreting human input. When users provide ambiguous, incomplete or erroneous information, the AI may misunderstand their intent or produce low-quality outputs.Input errors often reflect a mismatch between user expectations and system understanding. Addressing these gracefully is essential to maintain trust and ensure smooth interaction.How to use this patternHandle typos with grace: Use spell-checking or fuzzy matching to auto-correct common input errors when confidence is high (e.g., >80%), and subtly surface corrections (“Showing results for…”).Ask clarifying questions: When input is too vague or has multiple interpretations, prompt the user to provide missing context. In Conversation Design, these types of errors occur when the intent is defined but the entity is not clear. Know more about entity and intent. E.g., ChatGPT when given low-context prompts like “What’s the capital?”, it asks follow-up questions rather than guessing.Support quick correction: Make it easy for users to edit or override your interpretation. E.g., ChatGPT displays an edit button beside submitted prompts, enabling users to revise their input17. Design for AI system error statesGenAI outputs are inherently probabilistic and subject to errors ranging from hallucinations and bias to contextual misalignments.Unlike traditional systems, GenAI error states are hard to predict. Designing for these states requires transparency, recovery mechanisms and user agency. A well-designed error state can help users understand AI system boundaries and regain control.A Confusion matrix helps analyse AI system errors and provides insight into how well the model is performing by showing the counts of - True positives (correctly identifying a positive case) - False positives (incorrectly identifying a positive case) - True negatives (correctly identifying a negative case)- False negatives (failing to identify a negative case)Scenarios of AI errors and failure statesSystem failure (wrong output)False positives or false negatives occur due to poor data, biases or model hallucinations. E.g., Citibank financial fraud system displays a message “Unusual transaction. Your card is blocked. If it was you, please verify your identity”System limitation errors (no output)True negatives occur due to untrained use cases or gaps in knowledge. E.g., when an ODQA system is given a user input outside the trained dataset, throws the following error “Sorry, we don’t have enough information. Please try a different query!”Contextual errors (misunderstood output)True positives that confuse users due to poor explanations or conflicts with user expectations comes under contextual errors. E.g., when user logs in from a new device, gets locked out. AI responds: “Your login attempt was flagged for suspicious activity”How to use this patternCommunicate AI errors for various scenarios: Use phrases like “This may not be accurate”, “This seems like…” or surface confidence levels to help calibrate trust.Use pattern convey model confidence for low confidence outputs.Offer error recovery: Incase of System failure or Contextual errors, provide clear paths to override, retry or escalate the issue. E.g., Use way forwards like “Try a different query,” or “Let me refine that.” or “Contact Support”.Enable user feedback: Make it easy to report hallucinations or incorrect outputs. Read more about pattern 19. Design to capture user feedback.18. Design to capture user feedbackReal-world alignment needs direct user feedback to improve the model and thus the product. As people interact with AI systems, their behaviours shape and influence the outputs they receive in the future. Thus, creating a continuous feedback loop where both the system and user behaviour adapt over time. E.g., ChatGPT uses Reaction buttons and Comment boxes to collect user feedback.How to use this patternAccount for implicit feedback: Capture user actions such as skips, dismissals, edits, or interaction frequency. These passive signals provide valuable behavioral cues that can tune recommendations or surface patterns of disinterest.Ask for explicit feedback: Collect direct user input through thumbs-up/down, NPS rating widgets or quick surveys after actions. Use this to improve both model behavior and product fit.Communicate how feedback is used: Let users know how their feedback shapes future experiences. This increases trust and encourages ongoing contribution.19. Design for model evaluationRobust GenAI models require continuous evaluation during training as well as post-deployment. Evaluation ensures the model performs as intended, identify errors and hallucinations and aligns with user goals especially in high-stakes domains.How to use this patternThere are three key evaluation methods to improve ML systems.LLM based evaluations (LLM-as-a-judge) A separate language model acts as an automated judge. It can grade responses, explain its reasoning and assign labels like helpful/harmful or correct/incorrect.E.g., Amazon Bedrock uses the LLM-as-a-Judge approach to evaluate AI model outputs.A separate trusted LLM, like Claude 3 or Amazon Titan, automatically reviews and rates responses based on helpfulness, accuracy, relevance, and safety. For instance, two AI-generated replies to the same prompt are compared, and the judge model selects the better one.This automation reduces evaluation costs by up to 98% and speeds up model selection without relying on slow, expensive human reviews.Enable code-based evaluations: For structured tasks, use test suites or known outputs to validate model performance, especially for data processing, generation, or retrieval.Capture human evaluation: Integrate real-time UI mechanisms for users to label outputs as helpful, harmful, incorrect, or unclear. Read more about it in pattern 19. Design to capture user feedbackA hybrid approach of LLM-as-a-judge and human evaluation drastically boost accuracy to 99%.20. Design for AI guardrailsDesign for AI guardrails means building practises and principles in GenAI models to minimise harm, misinformation, toxic behaviour and biases. It is a critical consideration toProtect users and children from harmful language, made-up facts, biases or false information.Build trust and adoption: When users know the system avoids hate speech and misinformation, they feel safer and show willingness to use it often.Ethical compliance: New rules like the EU AI act demand safe AI design. Teams must meet these standards to stay legal and socially responsible.How to use this patternAnalyse and guide user inputs: If a prompt could lead to unsafe or sensitive content, guide users towards safer interactions. E.g., when Miko robot comes across profanity, it answers“I am not allowed to entertain such language”Filter outputs and moderate content: Use real-time moderation to detect and filter potentially harmful AI outputs, blocking or reframing them before they’re shown to the user. E.g., show a note like: “This response was modified to follow our safety guidelines.Use pro-active warnings: Subtly notify users when they approach sensitive or high stakes information. E.g., “This is informational advice and not a substitute for medical guidance.”Create strong user feedback: Make it easy for users to report unsafe, biased or hallucinated outputs to directly improve the AI over time through active learning loops. E.g., Instagram provides in-app option for users to report harm, bias or misinformation.Cross-validate critical information: For high-stakes domains (like healthcare, law, finance), back up AI-generated outputs with trusted databases to catch hallucinations. Refer pattern 10, Provide data sources.21. Communicate data privacy and controlsThis pattern ensures GenAI applications clearly convey how user data is collected, stored, processed and protected.GenAI systems often rely on sensitive, contextual, or behavioral data. Mishandling this data can lead to user distrust, legal risk or unintended misuse. Clear communication around privacy safeguards helps users feel safe, respected and in control. E.g., Slack AI clearly communicates that customer data remains owned and controlled by the customer and is not used to train Slack’s or any third-party AI modelsHow to use this patternShow transparency: When a GenAI feature accesses user data, display explanation of what’s being accessed and why.Design opt-in and opt-out flows: Allow users to easily toggle data sharing preferences.Enable data review and deletion: Allow users to view, download or delete their data history giving them ongoing control.ConclusionThese GenAI UX patterns are a starting point and represent the outcome of months of research, shaped directly and indirectly with insights from notable designers, researchers, and technologists across leading tech companies and the broader AI communites across Medium and Linkedin. I have done my best to cite and acknowledge contributors along the way but I’m sure I’ve missed many. If you see something that should be credited or expanded, please reach out.Moreover, these patterns are meant to grow and evolve as we learn more about creating AI that’s trustworthy and puts people first. If you’re a designer, researcher, or builder working with AI, take these patterns, challenge them, remix them and contribute your own. Also, please let me know in comments about your suggestions. If you would like to collaborate with me to further refine this, please reach out to me.20+ GenAI UX patterns, examples and implementation tactics was originally published in UX Collective on Medium, where people are continuing the conversation by highlighting and responding to this story.

Posted on : May 19, 2025 By Tech World Times Music  Rate this post You can enjoy your favorite music and audio content offline by converting YouTube videos to MP3 format, which frees you from the need for an internet connection. You can carry your music library with you wherever you go because MP3 files are smaller than video files and are therefore simpler to store and move between devices. Introduction YouTube offers a huge selection of music and audio content, so converting videos to MP3 format has grown in popularity. Converting YouTube videos to MP3 format can be a practical way to enjoy your favorite YouTube content offline, build a music library, or extract audio from podcasts. We’ll look at five quick and simple methods to convert YouTube videos to MP3 format in this essay. Method 1: Online Converters One common option for converting YouTube videos to MP3 format is to use an online converter. You can download the MP3 file from these websites by pasting the URL of the YouTube video. YouTubeMP3, FLVTO, and OnlineVideoConverter are a few well-known online converters. Method 2: Browser Extensions YouTube video conversion to MP3 format can be made easier with browser extensions. You can convert YouTube videos to MP3 format with just one click thanks to extensions like YouTube MP3, MP3 Converter, and Video Downloader Professional. Method 3: Desktop Software YouTube videos can be downloaded and converted to MP3 format using desktop software such as 4K Video Downloader, Freemake Video Converter, and Any Video Converter. These software applications frequently offer extra features like video editing and batch conversion. Method 4: Mobile Apps YouTube videos can be converted to MP3 format on your smartphone or tablet using mobile apps like YouTube MP3, MP3 Converter, and Video to MP3 Converter. These applications frequently offer a portable and easy way to convert YouTube videos. Method 5: Software with Built-in YouTube Support YouTube videos can be downloaded and converted to MP3 format using built-in features in some software applications, such as Audacity and VLC Media Player. Additional features like audio editing and format conversion are frequently offered by these programs. Selecting the Suitable Tools and Software for Conversion YouTube videos can be converted to MP3 using a variety of tools, including downloadable software and online converters. Selecting a tool that is dependable and easy to use is essential. To improve your conversion experience, look for features like batch conversion, multi-format support, and playlist download capability.Legal Considerations When Downloading YouTube Content It’s essential to understand the legal implications of downloading content from YouTube. Many videos are protected by copyright, and downloading them without permission may violate copyright laws. Always check the video’s description or consult the creator for permission if you plan to use the audio for anything beyond personal use, ensuring you respect the rights of content creators. Conclusion Online converters, browser extensions, desktop software, mobile apps, and software with integrated YouTube support are just a few of the quick and simple ways to convert YouTube videos to MP3 format. You can enjoy your favorite YouTube videos in MP3 format offline and on a variety of devices by following the detailed instructions provided above. Tech World TimesTech World Times (TWT), a global collective focusing on the latest tech news and trends in blockchain, Fintech, Development & Testing, AI and Startups. If you are looking for the guest post then contact at techworldtimes@gmail.com

Security is the cornerstone of many virtual private network (VPN) providers and one of the most common claims is operating and adhering to a no-logs policy.No-logs policies are public documents that decree the VPN provider will not hold, store, or sell customer data, including logs of their activities, timestamps, data transfers, or how they use VPN servers. If a VPN operates a no-logs policy and has independent audits to back it, this is considered an important measure in maintaining user privacy. While it's not possible to delve into VPN companies' inner infrastructure, a transparent approach to logs is a critical component of my recommendations. What is the best no-log VPN right now?ZDNET's top choice for a no-logs VPN and one most likely to protect your privacy is Mullvad. While one of the lesser-known VPNs around, Mullvad has proven how it handles user data requests and law enforcement requests with no data handed over. It has a fixed price of €5 ($5.60) per month. Our second favorite no-log VPN is NordVPN. NordVPN combines strong security, excellent speeds, and a vast server network with a no-logs policy and a transparent approach to privacy that many users trust. Plans start at $3.39 per month.Also: Best business VoIP providersZDNET writers have tested and reviewed VPNs for years. I rely on my experiences with the best VPNs on the market, frequent speed tests, and customer feedback to recommend the top VPNs you should consider before signing up. Below, you will find my selection of the best no-log VPNs. Sort by All The best no-log VPNs in 2025 Show less View now at Mullvad Mullvad VPN is a lesser-known VPN provider, but its customers know the organization takes privacy and security seriously, making it my top choice as a no-logs, open source VPN.Why we like it: Based in Sweden and protected by the country's strict data privacy laws, Mullvad operates a firm no-logs policy and does not collect or store activity logs. Payment data may be processed and protected by local laws, including General Data Protection Regulation (GDPR), but the company focuses on user anonymity. Mullvad goes so far as to number accounts rather than ask you for personal details such as your email address. Furthermore, I've seen how Mullvad deals with law enforcement requests. When subjected to a search warrant in 2023, Mullvad discussed the situation publicly and confirmed that while police intended to seize computers with customer data, no such devices existed for them to take. To back its transparent approach to security further, Mullvad allows users to verify its new Android app. The technical guidance is available on GitHub, alongside Mullvad's other open source projects.Who it's for: If you're looking for a VPN with a demonstrated approach to law enforcement requests and user data collection, check it out. I also like how easy this VPN is to use and its upfront pricing structure, although it would be great to see an expansion to its rather limited server network.  Mullvad offers a fixed pricing structure without changes or promotional deals. This way, you know whether you want a month, a year, or a decade of service, you will pay €5 ($5.60) per month. Who should look elsewhere: Customers like the transparent nature of this VPN and its provider, but would like to see more advanced features in the future. If you're looking for a large server network or a greater number of simultaneous connections, check out my next top pick. Mullvad VPN features: Simultaneous connections: 5 | Platforms: Windows, macOS, iOS, Android, Linux, Fedora Android TV, Fire TV, routers, and browser extensions | Countries: 49 | Kill switch: Yes | Logging: Numbered accounts, billing in accordance with law | Money-back guarantee: 14 days, some payment methods | Trial: No Pros Search warrant transparency Reasonable speeds Fixed pricing Open source Cons Limited simultaneous connections Low server network Mullvad VPN is a lesser-known VPN provider, but its customers know the organization takes privacy and security seriously, making it my top choice as a no-logs, open source VPN.Why we like it: Based in Sweden and protected by the country's strict data privacy laws, Mullvad operates a firm no-logs policy and does not collect or store activity logs. Payment data may be processed and protected by local laws, including General Data Protection Regulation (GDPR), but the company focuses on user anonymity. Mullvad goes so far as to number accounts rather than ask you for personal details such as your email address. Furthermore, I've seen how Mullvad deals with law enforcement requests. When subjected to a search warrant in 2023, Mullvad discussed the situation publicly and confirmed that while police intended to seize computers with customer data, no such devices existed for them to take. To back its transparent approach to security further, Mullvad allows users to verify its new Android app. The technical guidance is available on GitHub, alongside Mullvad's other open source projects.Who it's for: If you're looking for a VPN with a demonstrated approach to law enforcement requests and user data collection, check it out. I also like how easy this VPN is to use and its upfront pricing structure, although it would be great to see an expansion to its rather limited server network.  Mullvad offers a fixed pricing structure without changes or promotional deals. This way, you know whether you want a month, a year, or a decade of service, you will pay €5 ($5.60) per month. Who should look elsewhere: Customers like the transparent nature of this VPN and its provider, but would like to see more advanced features in the future. If you're looking for a large server network or a greater number of simultaneous connections, check out my next top pick. Mullvad VPN features: Simultaneous connections: 5 | Platforms: Windows, macOS, iOS, Android, Linux, Fedora Android TV, Fire TV, routers, and browser extensions | Countries: 49 | Kill switch: Yes | Logging: Numbered accounts, billing in accordance with law | Money-back guarantee: 14 days, some payment methods | Trial: No Read More Show Expert Take Show less Show less View now at NordVPN NordVPN is one of my top picks in the VPN market as a robust, speedy VPN with an expansive server network. Why we like it: NordVPN operates a strict no-logs policy that has been validated multiple times since 2018. Deloitte verified the policy's claims for the fifth time at the end of 2024. Furthermore, NordVPN is based in Panama, which does not enforce any data retention laws that would force the VPN provider to abandon this policy.While I would like to see support for unlimited simultaneous connections added in the future, NordVPN still stands out due to its excellent speeds, server network of over 110 countries, and multi-platform compatibility. Also: NordVPN reviewI've consistently found that NordVPN provides a stable, solid server network, fantastic speeds, and good connection times. Customers say this VPN is easy to use and that customer support is prompt and friendly.Additional security points of note include a double VPN for an extra layer of protection, a private domain name system (DNS) service, split tunneling, and meshnet, which operates like a secure local area network (LAN) and allows for secure data transfers.Who it's for: Anyone who wants a trustworthy VPN with a solid security background for all use cases. If you're interested in trying out NordVPN, the biggest savings can be found on two-year contracts, starting at $3.39 per month. If you agree, you will also be given free Sally eSIM data. Alternatively, one-year plans begin at $5 per month, or one month of access costs $13.Who should look elsewhere: NordVPN will suit many users as a no-logs VPN, but if you want the best speeds available, consider ExpressVPN.NordVPN features: Simultaneous connections: 10 | Platforms: Windows, macOS, iOS, Android, Linux, Chrome, Firefox, Edge, Android TV, Fire TV, tvOS, and routers | Countries: 118 | Kill switch: Yes | Logging: Email address and billing information only | Money-back guarantee: 30 days | Trial: Seven days on Android Pros Excellent speeds Massive server network Attractive user interface Independent no-logs verification Cons Expensive premium plans NordVPN is one of my top picks in the VPN market as a robust, speedy VPN with an expansive server network. Why we like it: NordVPN operates a strict no-logs policy that has been validated multiple times since 2018. Deloitte verified the policy's claims for the fifth time at the end of 2024. Furthermore, NordVPN is based in Panama, which does not enforce any data retention laws that would force the VPN provider to abandon this policy.While I would like to see support for unlimited simultaneous connections added in the future, NordVPN still stands out due to its excellent speeds, server network of over 110 countries, and multi-platform compatibility. Also: NordVPN reviewI've consistently found that NordVPN provides a stable, solid server network, fantastic speeds, and good connection times. Customers say this VPN is easy to use and that customer support is prompt and friendly.Additional security points of note include a double VPN for an extra layer of protection, a private domain name system (DNS) service, split tunneling, and meshnet, which operates like a secure local area network (LAN) and allows for secure data transfers.Who it's for: Anyone who wants a trustworthy VPN with a solid security background for all use cases. If you're interested in trying out NordVPN, the biggest savings can be found on two-year contracts, starting at $3.39 per month. If you agree, you will also be given free Sally eSIM data. Alternatively, one-year plans begin at $5 per month, or one month of access costs $13.Who should look elsewhere: NordVPN will suit many users as a no-logs VPN, but if you want the best speeds available, consider ExpressVPN.NordVPN features: Simultaneous connections: 10 | Platforms: Windows, macOS, iOS, Android, Linux, Chrome, Firefox, Edge, Android TV, Fire TV, tvOS, and routers | Countries: 118 | Kill switch: Yes | Logging: Email address and billing information only | Money-back guarantee: 30 days | Trial: Seven days on Android Read More Show Expert Take Show less Show less View now at ProtonVPN Protected by Swiss privacy laws, Proton VPN is another great VPN if you want a service that prioritizes security and privacy -- even on its free plan.Why we like it: Proton VPN operates a no-logs policy and does not collect or store data such as activity logs or session duration records. This VPN is also open-source, which means that any third party can examine its code for security or privacy flaws. Also: Proton VPNI've picked this VPN as one of the best options if you want a free plan, because while limited, it is backed by paid subscribers -- meaning you aren't trading your data for the service, despite not paying toward the running of the service.Free servers exist in five locations: the U.S., Romania, Poland, the Netherlands, and Japan. You are given one device connection.Who it's for: Whoever wants a free VPN plan. In general, customers like Proton's attitude to security and privacy, and many like the user interface. Although Trustpilot ratings are low, the feedback primarily relates to the firm's free plan limitations and customer support system. Despite this, as a free plan, Proton VPN is one of the few we recommend as a trustworthy service.Who should look elsewhere: If you're able to sign up for a paid option, I would do so -- as you will have full access to Proton VPN's network and better speeds, as you won't be limited to a small subset of server locations.Proton VPN's pricing is reasonable, but not as cheap as some of its competitors. Plans lasting two years offer the best value for money and are priced at $4.49 per month. Alternatively, annual plans cost $5 per month, whereas one month of service is $10.Proton VPN features: Open-source | Simultaneous connections: One to 10, depending on plan | Countries: 117 (5 free) | Logging: No-logs policy | Security audits | Platforms: Windows, macOS, iOS, Android, Linux, Chrome, Firefox, Edge, Android TV, Fire TV, Kodi, and routers | VPN accelerator | Money-back guarantee: 30 days | Trial: Free option Pros Security pedigree Open source Free plan Cons Limited simultaneous connections Protected by Swiss privacy laws, Proton VPN is another great VPN if you want a service that prioritizes security and privacy -- even on its free plan.Why we like it: Proton VPN operates a no-logs policy and does not collect or store data such as activity logs or session duration records. This VPN is also open-source, which means that any third party can examine its code for security or privacy flaws. Also: Proton VPNI've picked this VPN as one of the best options if you want a free plan, because while limited, it is backed by paid subscribers -- meaning you aren't trading your data for the service, despite not paying toward the running of the service.Free servers exist in five locations: the U.S., Romania, Poland, the Netherlands, and Japan. You are given one device connection.Who it's for: Whoever wants a free VPN plan. In general, customers like Proton's attitude to security and privacy, and many like the user interface. Although Trustpilot ratings are low, the feedback primarily relates to the firm's free plan limitations and customer support system. Despite this, as a free plan, Proton VPN is one of the few we recommend as a trustworthy service.Who should look elsewhere: If you're able to sign up for a paid option, I would do so -- as you will have full access to Proton VPN's network and better speeds, as you won't be limited to a small subset of server locations.Proton VPN's pricing is reasonable, but not as cheap as some of its competitors. Plans lasting two years offer the best value for money and are priced at $4.49 per month. Alternatively, annual plans cost $5 per month, whereas one month of service is $10.Proton VPN features: Open-source | Simultaneous connections: One to 10, depending on plan | Countries: 117 (5 free) | Logging: No-logs policy | Security audits | Platforms: Windows, macOS, iOS, Android, Linux, Chrome, Firefox, Edge, Android TV, Fire TV, Kodi, and routers | VPN accelerator | Money-back guarantee: 30 days | Trial: Free option Read More Show Expert Take Show less Show less View now at PIA Private Internet Access, also known as PIA, is another of our favorite no-logs VPNs. Why we like it: This service is open source and its stringent privacy policy is frequently audited by independent parties, including Deloitte, verifying that PIA does not collect or store your data. PIA also makes use of RAM-only servers, which removes traces of online activity. Review: Private Internet AccessDuring our speed tests, PIA performed reasonably well, and we like the fact that this VPN provides unlimited simultaneous connections for its price point. Customers, in particular, applaud its Linux support.Who it's for: Anyone who wants a consistently affordable option with unlimited simultaneous connections.The best deal right now is for three-year contracts, which cost only $2.19 per month and come with two extra months of service. You can also sign up for six months, priced at $7.50 per month.Who should look elsewhere: If you want a faster VPN with a larger server network and a similar security pedigree, I'd take a look at NordVPN. It will cost a little more, though. Private Internet Access features: Simultaneous connections: Unlimited | Kill switch: Yes | Platforms: Windows, macOS, iOS, Android, Linux, Chrome, Firefox, Edge, tvOS, routers, Fire TV, and NAS | Countries: 91 | Logging: No-logs policy | Money-back guarantee: 30 days | Trial: Seven days, Android and iOS Pros Unlimited simultaneous connections Mobile trials Open source Six-month contract available Cons Average speeds Limited server network No yearly contract Private Internet Access, also known as PIA, is another of our favorite no-logs VPNs. Why we like it: This service is open source and its stringent privacy policy is frequently audited by independent parties, including Deloitte, verifying that PIA does not collect or store your data. PIA also makes use of RAM-only servers, which removes traces of online activity. Review: Private Internet AccessDuring our speed tests, PIA performed reasonably well, and we like the fact that this VPN provides unlimited simultaneous connections for its price point. Customers, in particular, applaud its Linux support.Who it's for: Anyone who wants a consistently affordable option with unlimited simultaneous connections.The best deal right now is for three-year contracts, which cost only $2.19 per month and come with two extra months of service. You can also sign up for six months, priced at $7.50 per month.Who should look elsewhere: If you want a faster VPN with a larger server network and a similar security pedigree, I'd take a look at NordVPN. It will cost a little more, though. Private Internet Access features: Simultaneous connections: Unlimited | Kill switch: Yes | Platforms: Windows, macOS, iOS, Android, Linux, Chrome, Firefox, Edge, tvOS, routers, Fire TV, and NAS | Countries: 91 | Logging: No-logs policy | Money-back guarantee: 30 days | Trial: Seven days, Android and iOS Read More Show Expert Take Show less Show less View now at ExpressVPN If speed and a strict no-logs policy is what you're after, ExpressVPN will be the right choice for you.Why we like it: ExpressVPN also operates a strict no-logs policy and will not collect or store personal data, including browsing history, traffic records, domain name service (DNS) queries, or internet protocol (IP) addresses. The company uses RAM-only servers for additional data protection, and independent audits of its no-logs policy have been conducted since 2019.Also: ExpressVPN reviewWe like this VPN's extreme speed. It always performs very well on our speed tests, taking center stage as the best on offer, with a robust 105-country server network as a bonus.Who it's for: Those looking for the fastest no-logs VPN on the market today. It's fast, user-friendly, and in my experience, both connection and network speeds are some of the best available.Who should look elsewhere: We would like to see the price reduced, and this sentiment is reflected in customer reviews. It has dropped from an average of $6.67 per month since last year, but pricing is not yet in line with many competitors. It is an excellent service, but not the cheapest. The two-year subscription -- with four extra months included -- costs $5 per month. The one-year plan costs $6.67 per month, and one month of service is $12.95. A 7-day free trial is included with all plans and works on every supported device.ExpressVPN features: Simultaneous connections: 8 | Kill switch: Yes | Platforms: Windows, macOS, iOS, Android, Chrome, Firefox, Edge, Linux, Fire TV, Android TV, tvOS, and routers | Countries: 105 | Logging: No browsing logs, some connection logs | Money-back guarantee: 30 days | Trial: Seven days Pros Excellent speeds Large server network Mobile trials User-friendly Cons Expensive No unlimited simultaneous connections If speed and a strict no-logs policy is what you're after, ExpressVPN will be the right choice for you.Why we like it: ExpressVPN also operates a strict no-logs policy and will not collect or store personal data, including browsing history, traffic records, domain name service (DNS) queries, or internet protocol (IP) addresses. The company uses RAM-only servers for additional data protection, and independent audits of its no-logs policy have been conducted since 2019.Also: ExpressVPN reviewWe like this VPN's extreme speed. It always performs very well on our speed tests, taking center stage as the best on offer, with a robust 105-country server network as a bonus.Who it's for: Those looking for the fastest no-logs VPN on the market today. It's fast, user-friendly, and in my experience, both connection and network speeds are some of the best available.Who should look elsewhere: We would like to see the price reduced, and this sentiment is reflected in customer reviews. It has dropped from an average of $6.67 per month since last year, but pricing is not yet in line with many competitors. It is an excellent service, but not the cheapest. The two-year subscription -- with four extra months included -- costs $5 per month. The one-year plan costs $6.67 per month, and one month of service is $12.95. A 7-day free trial is included with all plans and works on every supported device.ExpressVPN features: Simultaneous connections: 8 | Kill switch: Yes | Platforms: Windows, macOS, iOS, Android, Chrome, Firefox, Edge, Linux, Fire TV, Android TV, tvOS, and routers | Countries: 105 | Logging: No browsing logs, some connection logs | Money-back guarantee: 30 days | Trial: Seven days Read More Show Expert Take Show less No-log VPNs, compared No-log VPNNo-logs policyWarrant canary/transparency reports?Server networkPriceMullvad VPNYesNews publications49$5.60 per monthNordVPNYesTransparency reports118$3.39 per monthProton VPNYesYes, plus transparency reports117$4.49 per monthPrivate Internet AccessYesTransparency reports91$2.03 per monthExpressVPNYesTransparency reports105$5 per month Show more What is a no-log VPN service and why should you use one? A VPN service is a virtual private network that encrypts your internet traffic, hides your IP address, and provides a relatively anonymous and secure connection. A VPN makes it more difficult for people to digitally eavesdrop or monitor your online activities, but there is no point if your information is being collected or exposed. That's why choosing a no-logs VPN is crucial. Otherwise, some free VPN providers offer a VPN in exchange for your data and records. That's why you should pick a paid plan from a reputable VPN with a no-logs policy.See also: What is a VPN and why do you need one? Show more How to set up a no-log VPNOnce you've chosen the VPN you want to install, head to an official software repository -- such as Apple's App Store or the Google Play Store -- or directly to the VPN's website, which should have links to all the download sources. June Wan/ZDNETDownloaded the app and install it. You may have to provide the VPN with specific device permissions. Once the software is installed, open it and log in with your credentials, QR code, or one-time passcode. Also: Best password generators of 2025 ZDNETSelect a country or choose quick connect, and you're done.See also: How to set up a VPN on your routerHow we test VPNs Many providers recommended in this list have been subject to in-depth testing and reviews by the ZDNET team. We've been talking about VPNs for years, spoken with their management and their users, and developed a generally favorable impression. ZDNET tests VPN services from multiple locations, but we can't test from all locations. Every home, community, local internet service provider (ISP), and nation has a different infrastructure. I analyzed the no-log best VPNs below using the following criteria that are most important in helping you make an informed decision: No-logs policies: We thoroughly examined the transparency and security standards of each VPN we recommend. Each VPN must operate a no-logs policy and keep user data collection to a minimum.Security: There's no point in using a VPN unless its security protocols and encryption levels are up to scratch. We also ensured the development team behind each recommended VPN cares about making security improvements and patching reported bugs.Performance: Our top VPNs must have a minimum level of performance. After all, there's no point in using a VPN if your connection continually drops or lags.Simultaneous connections: Many VPNs limit the number of devices you can connect at the same time. If you have a phone, tablet, laptop, desktop, or smart TV you want to connect to a VPN, you need a service that allows at least five or six simultaneous connections.Kill switch: A kill switch prevents you from having an unprotected connection. It is an important feature for a VPN, as it protects you by preventing a return to a default Wi-Fi connection should your internet connection unexpectedly drop.Geo-blocks and streaming: VPNs can be key to accessing local content and services while you're away from home. We ensured that our top recommendations performed well given these factors and are unlikely to slow down your connection while streaming. Platforms: What systems and devices can you run the VPN on? Our no-logs VPN recommendations naturally have to offer easy-to-use and intuitive apps.Support: As with any subscription software, you may encounter problems with installing apps, using the service, renewing, or canceling. We list VPNs with an available support team.Countries and servers: In how many different countries does the VPN have servers? This can include physical and virtual servers, and whether there are enough -- at least 50 or so -- to provide variety.Trial length and price: Each VPN we recommend has a solid money-back guarantee.For a more extensive breakdown, check out our comprehensive VPN testing methodology page. Show more Choosing the right no-logs VPN Choose this VPN...If you want...Mullvad VPNA transparent no-logs VPN. Mullvad is based in Sweden and subject to strict data protection laws. The company is transparent in regard to police activities.NordVPNThe best overall no-logs VPN. NordVPN provides the whole package: robust security, speed, and a validated no-logs policy.Proton VPNA free VPN focused on security and privacy. While the free plan is limited, it is backed by paid subscribers, so your data should not be collected and sold.Private Internet AccessA no-logs VPN with unlimited connections. PIA is open-source, uses RAM-only servers, and is known for its stringent approach to security.ExpressVPNA speedy no-logs VPN. ExpressVPN provides fantastic speeds and is easy to use, and audits have been conducted to verify its no-logs policy.  Show more What are the factors to consider when choosing a no-logs VPN? When you are selecting your new no-logs VPN, consider the following factors:Privacy: It is important that your new VPN takes user privacy and security seriously. Look for no-logs pledges, transparent security reports, warrant canaries, and practices like frequent patch updates.Price: You need to consider the overall price of your new VPN plan. Try to take advantage of promotions, but ignore countdown timers -- they typically are just marketing tools. While top-notch VPNs can cost upwards of $10 per month, many are available for only a few dollars. Many VPN providers will also offer free, limited mobile app trials. We recommend that you avoid free VPNs -- with a few exceptions -- as they may sell your data or bombard you with ads.Simultaneous connections: If you're paying for a VPN, you might want to use it across multiple devices, such as a desktop, a laptop, or smartphone. We recommend that you sign up for a plan supporting at least six devices. Thankfully, many paid VPN plans now offer unlimited connections.Contract length: You must also decide how long you want to use a VPN. If you opt for a two-year term, you tend to secure the best prices, but other options are available. After you've made a selection, you typically have a 30-day money-back guarantee.Apps: When selecting a VPN, you should consider what devices you want to protect. For example, if you have a Windows PC and an iPhone, take a look at the apps on offer to see if you like their user interfaces and functions. Show more FAQs What does no-logs mean? A no-logs policy ensures that a VPN will not log your activity including the websites you visit, the online services you access, or domain name service (DNS) queries when your requests are sent through its server network. A no-logs policy doesn't mean that the VPN provider doesn't log all traffic and data. For example, a VPN provider may still log your VPN usage records, sessions, and billing information. They may also be required to log some customer information to comply with the laws of the country they are headquartered in. Show more Can the police track a VPN? The police can request user information from a VPN provider in the same way they do for any other organization. With a court order, the VPN vendor will have to comply, but the information they can provide depends on their logging activities.Law enforcement may be able to detect VPN use and may investigate IP addresses associated with you, but it's highly unlikely they would go down this route unless you are suspected of involvement in illegal activity. Remember, a VPN is not a foolproof privacy barrier and does not guarantee full anonymity.  Show more How do I tell if my VPN is safe? To check if your VPN is safe and working properly, first choose a reputable vendor known to have a strong security background.Next, compare your IP address before and after enabling a VPN connection. You can find this under your device's settings, typically under Wi-Fi, or by visiting whatismyipaddress.com. If they both display the same IP address, your VPN may not be working properly.You can also visit dnsleaktest.com to see if any domain name service (DNS) leaks are detected while you are using a VPN, which could be a sign that your information is being leaked.Finally, be careful with free VPNs offered by unknown or shady companies, especially if they are outside of official app repositories. This could open your device to surveillance or malware.  Show more Are no-logs VPN services legal? Yes, no-logs VPN services are legal in the U.S., but the legality of using VPNs varies from country to country.It is important to check the laws of your location before using a VPN. Some countries ban their use outright such as North Korea, Belarus, and Iraq, whereas other countries including Russia and China, frown upon using any non-state-controlled VPN.For a list of countries where VPN usage is illegal, check out NordVPN's guide. Show more Can my employer see my browsing history if I use a VPN? Potentially, yes. If your VPN is working properly, your data and activities are hidden within an encrypted tunnel, and you have improved protection and anonymity online, but this doesn't mean that your browsing history isn't logged through your browser or that monitoring software on corporate PCs isn't logging your activities. Corporate VPNs may also log your website visits.  Show more Are there alternative no-logs VPNs to consider?Our recommendations above are the best no-logs-focused VPNs on the market today. If they aren't quite right for your particular needs, there are other trustworthy options below.Latest updates In ZDNET's May update, we performed substantial copy and layout changes. We also featured Mullvad as our top pick. Show more Further ZDNET Tech Coverage Smartphones Smartwatches Tablets Laptops TVs Other Tech Resources ZDNET Recommends