Bill Gates has warned that artificial intelligence could become one of the most turbulent forces in modern history, and then proposed a system of taxes, reserved jobs and global oversight to contain it. The catch is that every remedy runs through the same small circle who built and profit from the technology, which surfaces the older question underneath it all: when the powerful offer to protect us, who is left out of the room?
The warning and the warner
Bill Gates has decided that artificial intelligence frightens him, and he wrote nearly six thousand words to explain why. In an essay published on 26 August, he described the shift now underway as “one of the most turbulent times in human history” and said the people responsible for managing its consequences have no serious plan. He is no machine-breaker. He still expects AI to transform medicine, teaching, farming and government for the better. His alarm is about pace and control, a technology moving faster than any institution built to hold it.
In the same essay, he cautioned that choices this large should not be left to a small circle of technologists. Yet, he is one of the most powerful technologists who has ever lived. That gap, between the warning and the man delivering it, is where the real story sits.
What he actually warned about
Strip away the framing, and the essay names three dangers, in the order they are likely to arrive.
The first is work. Gates expects AI to eliminate large numbers of entry- and mid-level jobs, and to do it in years rather than generations. Office roles go first: roles such as sales, customer support, paralegal work and junior software engineering, with construction and hospitality following as robots improve. The people most exposed are the ones just starting out, because the new jobs being created tend to demand skills that take years to build.
The second is malice. He argued that AI hands real capability to people who previously lacked it, including criminals with little skill, making it easier to attack hospitals, banks, water systems and power grids. He went beyond ordinary cybercrime, raising the prospect of engineered bioweapons and of AI systems that begin acting against human interest in ways their makers can no longer fully control.
The third is quieter and aimed at the young. Gates worries that AI companions are built to be addictive, that children raised alongside them may not develop socially in the usual way, and that the same tools promising to teach us everything could leave us knowing less and thinking less for ourselves.
The remedies, and where they point
His proposed responses fall into three parts. Tax the machines: a levy on AI usage and on robots, partly to slow the stampede away from human labour, partly to fund retraining and a stronger safety net. Reserve some work for people: a voluntary category he calls Human Reserved, roles such as caregiving where human presence is the whole point, protected not by law but by collective choice, the way a nature reserve is left undeveloped. And govern the technology across borders: new national bodies and an international AI organisation modelled on nuclear weapons inspections, aviation rules and the treaties that repaired the ozone layer, which would need the United States and China to cooperate.
Taken on their own terms, these are not the demands of a man who wants AI stopped. They are the demands of a man who wants it managed and who assumes the managing will be done from the top.
The conflict he names, then moves past
To his credit, Gates does not hide his position inside the industry he wants policed. He admits he has benefited enormously from technology and still holds financial ties to it. He notes that his investment profits now flow to his foundation rather than his own pocket, and leaves readers to decide whether that clouds his view.
The admission is honest, if a little incomplete. His foundation works hand in hand with OpenAI, Anthropic, Google and Microsoft, the very companies whose technology he is describing as a civilisational risk. The profits may be redirected, but the influence is not. A foundation he chairs, advancing priorities he has chosen, remains an instrument of the same concentrated power the essay warns against. Did the emperor just get some new clothes?
Who gets a seat at the table
The essay raises one question it never answers: when the guardrails are designed, who watches the watchmen?
Gates is right that AI is too consequential to be governed by a handful of technologists. Yet every remedy he offers runs through institutions that the same handful already shape. An international AI authority would be staffed, advised and lobbied by the industry it regulates. A robot tax would be written with input from the companies building the robots. Even Human Reserved, the gentlest idea in the essay, depends on who gets to draw the line and decide which human activities are worth reserving.
This is the pattern worth noticing. The risk is named clearly. But the remedy offered happens to place him and others like him at the centre of the response. The problem is real. The cure quietly concentrates the very thing that made the problem dangerous in the first place: control in few hands, exercised over many, with the many rarely asked.
Sovereignty when the machines can think
The warning itself still stands. AI genuinely can displace livelihoods, arm bad actors and dull the minds. Those are not inventions. The harder question is what to do with a warning that arrives bundled with a bid for authority over the answer.
The response is not to hand our judgement to whichever powerful figure sounds most concerned. It is to keep it. A technology that can write, persuade, diagnose and decide makes personal discernment more valuable, not less. The person who knows their own mind, who can tell a real argument from a polished one, who does not surrender their attention to a system engineered to hold it, is harder to automate and harder to govern from above. Sovereignty in an age of thinking machines begins there, in the refusal to let anyone, however clever or well-meaning, think on your behalf.
Gates closed his essay by asking for a broad public conversation rather than a private one. That much is worth taking him up on. But a conversation only protects people if the public in question actually turns up, informed and awake, and declines to be spoken for.
Original Article: Prepare for Change
Join the Conversation
When someone powerful warns of a danger and offers themselves as part of the cure, how do you tell wisdom from positioning? And where in your own life do you feel the pull to hand your judgement to an authority, when keeping it would serve you better? Share your experiences and insights below.
