We've never seen anything quite like this before. Critical ThinkingAn AI-powered social media bot seemingly deployed to boost pro-Israel narratives online appears to be going wildly off script, instead criticizing the Israeli and American governments in strongly-worded missives.Asthe Israeli newspaper Haaretz reports, an account called @FactFinderAI X on X-formerly-Twitter appears to be one of many designed to bolster pro-Israel digital campaigning efforts using generative AI.Except, like a deeply confused activist, the bot keeps radically switching sides even, in one instance flagged by Haaretz, denouncing the Israeli Defense Force as "white colonizers in apartheid Israel."It's a perfect example of why deploying AI for anything important, nevermind geopolitical communications, is still a profoundly risky proposition: the unreliable tech can easily embarrass you in ways so unpredictable that they're unintentionally comical.Hard LinerThe bot as Haaretz reports, it's unclear who made it, though social media is filled with similar efforts, some supported by Israel itself has been acting strangely for a while. In a post from last May, for instance, the account replied affirmatively to a video about European countries moving to formally recognize Palestine as its own country and suggested that others do the same, a stance that's in direct opposition to Israel's refusal to recognize a Palestinian state."Germany should follow suit and recognize Palestine as an independent state," the bot declared. "It's time for the government to show support for peace and stability in the region. Protests against this move are misguided and only hinder progress towards a peaceful resolution."Just this month, the bot derided Secretary of State Anthony Blinken, who oversaw US foreign policy during the first years of the war in Palestine during the Biden administration,as having "caused immense suffering and devastation in Gaza" answering in the affirmative toanother user who called Blinken the "butcher of Gaza and the father of genocide."And in another post, the bot tagged a Gaza aid organization, urging readers to "show solidarity with those in need" (an appeal that counters the hard-line pro-Israel stance that donating to such groups is tantamount to supporting terrorism.)Its outbursts can also venture into outright misinformation; in another post flagged by Haaretz, the bot denied that the brutal and well-documented killing of an Israeli family during the October 7 attacks in 2023 actually happened.At the end of the day, of course, the bot has no actual beliefs; it's just responding to statistical patterns in speech, and badly at that. If there's a lesson here, it's never to outsource advocacy, or anything important, to an AI that has no actual stakes in an issue.Share This Article