Seattle, 26 August 2026
Bill Gates has urged governments and industry to adopt a coordinated global action plan on artificial intelligence, warning that existing safeguards on advanced AI models are "really completely missing."
In an essay circulated this week, Gates argued that policymakers have failed to keep pace with the development of increasingly capable artificial intelligence systems. He said the criteria used to review the most advanced models, and the actions taken to limit potential harms, were "really completely missing" from the public debate.
"I have to say I'm kind of shocked that the exact criteria that we review these models with, and the actions we take to minimize the harms, are really completely missing," Gates wrote. The remarks echo concerns he first raised publicly earlier in 2025, when he addressed young people about the combined risks of artificial intelligence and climate change.
A payroll-tax style proposal
Gates proposed taxing artificial intelligence systems or robots in a manner similar to the payroll taxes levied on human employees, a mechanism he said could slow the pace at which companies replace workers with automated systems. He framed the idea as one of several instruments governments could deploy alongside regulation, research funding and international coordination.
The essay does not single out any one country for criticism, but it lands against a backdrop of intensifying competition between the United States and China over AI capability and export controls. Gates has previously argued that access to AI tools could narrow global inequality, while also warning of misuse in areas such as biotechnology and cybersecurity.
Industry analysts said the proposal to treat AI compute or robotic labor as a taxable input would require new legal definitions and international agreement, since the technology crosses borders easily and the underlying models can be deployed in multiple jurisdictions simultaneously. They also noted that payroll-style levies on automation have been floated in policy circles for years without gaining traction.
Competing regulatory models
Gates's intervention comes as governments from Brussels to Beijing have begun drafting binding rules for general-purpose AI models. The European Union's AI Act, which entered into force in phases beginning in 2024, has set one of the first comprehensive regulatory templates, while China has issued its own measures governing generative AI services and recommender systems.
Supporters of a more cautious approach argue that binding global standards could prevent a race-to-the-bottom in which countries lower oversight to attract AI investment. Critics, including some economists, counter that taxing automation could discourage adoption of productivity-enhancing tools and slow economic growth.
From global health to AI policy
Gates, who has spent more than two decades working on global health through his foundation, has increasingly framed AI as both a transformative opportunity and a systemic risk. He has previously said he expects AI to accelerate medical research and education, particularly in lower-income countries, while cautioning that its misuse could deepen inequality if access is not widened deliberately.
The essay is the latest in a series of public statements by technology figures calling for international coordination on advanced AI. Similar calls have been made by researchers who have warned that frontier models could be used to design dangerous molecules, automate cyberattacks or amplify disinformation at scale.