Mark Zuckerberg’s AI manifesto illustrates why the public distrusts artificial intelligence

Mark Zuckerberg published this week a 6,500-word manifesto on personal artificial intelligence, the most detailed version to date of Meta AI’s vision for “personal super-intelligence.” The ideas are not new; a shortened version ran in the Wall Street Journal two weeks ago, and Zuckerberg has been talking about them in investor calls for a while, but this time he laid out the full picture, and that is precisely the problem.

The text is sometimes described as an “anti-dumb” article, but that is inaccurate. It is more a description of why Zuckerberg himself is excited about the ways AI will change society. The problem: while painting a rosy future, he repeatedly mentions all the ways it could go wrong. In the background the social-media crisis looms: a recent poll found that “64% of Americans believe social media is linked to democracy”, and that “a similar share supports stricter regulation”, figures that cut across party lines. Just this weekend a court fined the company 567 million dollars for harm to children. The atmosphere, to put it mildly, is not positive.

A large portion of the manifesto is devoted to vague generalities about intelligence, reminiscent of the broad statements Zuckerberg and Jack Dorsey once made about free speech. For example: “The more powerful tools everyone has, the more each person will be able to shape the future, no less. People and institutions with competing interests balance each other naturally and lead to positive outcomes.” It sounds nice, but one can think of many cases where competing interests produced worse outcomes. More oddly, it is presented as an earnest philosophical conclusion rather than a justification for a specific product called “personal intelligence” that Meta is launching. Readers already inclined not to trust the author may suspect he wrote it mainly to convince himself that nothing bad will happen.

The feeling intensifies when he reaches concrete examples, some of which are disturbingly detached from how AI tools are actually used. Take education: Zuckerberg describes a personalized private tutor with a doctorate in every subject and unlimited patience, helping students in places where today only wealthy parents can afford assistance, and providing adults with superhuman learning aides for skills, languages and hobbies. The product he describes already exists: it is a consumer chatbot like ChatGPT, Claude and Gemini. The promise of a “doctorate in every subject” is exactly the kind of hype the public has learned to spot and reject.