Bill Gates has published a new memo calling for an artificial intelligence tax, a reserved category of jobs for human workers, and new institutions to manage the transition. In an interview published this week, he said that the people building AI are worried in private and quiet in public: 'I know they're all worried. Or all of them that I know, which is basically everybody but Elon.'
Key facts from the announcement
- Bill Gates proposes an AI tax, a human-only job category, and new institutions to manage the transition.
- He says most AI industry leaders are privately worried and publicly silent, with Elon Musk the exception.
- Gates first proposed a robot tax in February 2017, the same month the European Parliament voted on robotics rules without including the tax.
- He warns that even lack of control he expected years from now is already visible.
Three proposals in the memo
Gates argues that the industry is crossing safety lines it set for itself, and that no one is preparing for what follows. The memo contains three main ideas. First, an AI tax that would slow the deployment of certain systems and generate revenue for workers affected by automation. Second, a category of jobs reserved exclusively for humans, ensuring that some forms of employment remain outside the reach of machines. Third, new institutions designed to oversee the transition and provide guidance to governments and companies.
In the interview, Gates went further than the written memo. 'Even lack of control that I thought would be many years from now, we're seeing lack of control,' he said. That admission is striking because Gates has often described AI as the most important technological advance in decades, sometimes comparing it to the birth of the internet.
His position on taxation is not new. In February 2017, he first proposed taxing robots when the debate centered on factory automation rather than generative models. The reasoning was the same: automation would eliminate jobs faster than new ones could be created, and the companies benefiting from that automation should contribute to the social safety net.
Europe answered that question the same month. The European Parliament voted 396 to 123, with 85 abstentions, on a resolution setting out civil law rules for robotics. The tax was not included in the final text. Instead, the resolution asked the European Commission to examine whether member states' social security systems could remain viable as robots spread. It also floated the idea of electronic personhood for the most sophisticated autonomous robots. That concept has not gained traction in the nine years since.
The debate in 2017
The 2017 Parliament resolution was an early attempt to address the rise of robotics and artificial intelligence. Mady Delvaux, the rapporteur, wanted a forward-looking debate about the future of work. She later said that certain coalitions in Parliament had blocked that debate, while the robotics industry welcomed the outcome. The compromise left Europe with a set of principles but no fiscal mechanism and no clear legal status for intelligent machines.
Electronic personhood was one of the most controversial ideas. The idea was to give advanced autonomous systems a legal identity similar to the one corporations have. Proponents argued that it would make it possible to attribute responsibility when an autonomous system caused harm. Opponents worried that it would grant rights to machines and reduce the accountability of manufacturers and operators. The concept was widely criticized by academics, labor unions, and even the European Commission's own advisers, and it has not reappeared in any major legislative proposal.
The notion of a robot tax also failed to gain support. Economists remain divided. Some argue that taxing automation would slow productivity and discourage innovation. Others believe that if labor is taxed and capital is not, governments are effectively subsidizing the replacement of workers. A robot tax, in this view, is simply a way to level the playing field.
Nine years later
Nine years after the 2017 vote, the European Union has enacted the AI Act, a comprehensive regulation that takes a risk-based approach to artificial intelligence. It imposes obligations on providers and users, bans certain uses of AI, and requires human oversight for high-risk systems. But it does not do what Gates has proposed. It does not tax models. It does not reserve a category of work for humans. And no member state has introduced a national robot tax or a human-only job category.
The AI Act is designed to regulate the behavior of AI systems, not to manage the broader economic transition. It includes provisions on transparency, data quality, and accountability, and it creates a governance structure involving national authorities and a new European AI Office. But its focus is safety and fundamental rights, not labor market policy. That gap may explain why Gates is repeating his proposal now