Zuckerberg Wrote 6,500 Words to Calm Us Down About AI: In Every Other Example He Explains Why We're Worried

Mark Zuckerberg has published a 6,500-word essay on how "personal superintelligence" will change every person's life. A version of the text already ran in the Wall Street Journal two weeks ago, and he repeats the same ideas on quarterly investor calls. This is simply the most detailed repackaging so far - and the details are precisely the problem.

The text is conceived as an answer to the fears. Instead, every few paragraphs it reminds you why the fears exist. When a man trying to convince you there is nothing to be afraid of keeps citing examples in which things could go wrong, you don't calm down - the opposite.

One part of the essay revolves around everyone getting more powerful tools, so power will level out by itself. "People and institutions with opposing interests naturally check and balance each other," Zuckerberg writes. The Balkans know this sentence by heart - it is the same logic used to explain that the free market would sort things out on its own. Opposing interests sometimes balance. Sometimes one of them simply wins.

The weakest point is the education example. Zuckerberg promises everyone a personal tutor with a doctorate in every field and unlimited patience. That product already exists - it is ChatGPT, Claude and Gemini. Except that the main use of those tools in the classroom isn't learning but avoiding learning: the machine writes the homework, and the teacher has no reliable way to check which assignment a pupil actually wrote. That isn't a hypothetical future risk. That is what is happening this semester.

The problem is trust, not technology

The numbers behind this mood are uncomfortable for Meta. According to a recent survey, 64 percent of Americans believe social media has damaged democracy, and a similar share wants stricter regulation - equally among voters of both parties. The same weekend the essay came out, a court fined the company 567 million dollars (around 490 million euros) over harm to children.

That is the real story. Anxiety about artificial intelligence didn't fall out of the sky - it is the direct inheritance of what happened with social media. The public already believed the tech industry once that a new product would improve society. An essay that skips those consequences and again promises only good doesn't restore trust; it demonstrates how the trust was lost.

Even the sections about his own business turn oddly abstract. Zuckerberg says everyone will have free or affordable access to the tools, with free versions for billions of people. In principle that's true - a free model does open wider access than everyone paying up front for compute. But there is a reason no consumer product charges the user the live market price of servers: billing that spikes with demand is a catastrophic experience when you need the tool for work.

The comparison with the competition doesn't favour Meta. Sam Altman and Dario Amodei, whatever their flaws, at least do what Zuckerberg refuses here - they acknowledge the technology carries dangers, explain what measures they are taking and try to convince you they can be trusted. That is no guarantee of anything. But when the safeguards give way, and sooner or later they do, an industry with no accumulated trust has nothing to lean on.

For a reader in this region the question is simpler. If the man building the tool sees no need to talk about the risks, who will talk about them when they reach us?