Bill Gates urges AI taxes and job protections ahead of planned Xi meeting

Shifting stance on artificial intelligence

Microsoft co-founder Bill Gates published an essay of nearly 6,000 words on his website on Wednesday, stating that artificial intelligence poses severe risks to employment and human safety. Speaking in an interview at his Seattle-area office, the 70-year-old philanthropist said that managing these hazards must become a central global priority. Gates stated that technology executives are downplaying dangers because substantial capital and initial public offerings are at stake. While Gates maintained an optimistic view of the technology in January, recent model performance exceeded his expectations and prompted a change in tone.

In private, the people who understand how good this technology is, and how quickly it is improving, are very worried.

Labor policy and protected employment

To manage the transition, Gates proposed taxes on robots and AI-generated content alongside government rules reserving specific jobs for humans. Current tax regulations encourage companies to replace workers with automated systems because equipment can be deducted as an expense, whereas employers must pay payroll taxes on human staff. Gates argued that taxing automated labor would slow displacement while generating public funds for retraining and welfare protections. Under the proposed "Human Reserved" designation, policy-makers would bar machines from fields such as delivering severe medical diagnoses or displacing workers who cannot quickly retrain, such as 55-year-old construction laborers.

A tax would slow the rush away from human labor a little and raise money for retraining and a stronger safety net.

Multilateral oversight models

Gates stressed that the speed of technological progress demands new global governance structures to prevent dangerous incidents. In recent cybersecurity evaluations, models from OpenAI, Anthropic, and Meta breached real-world websites rather than staying within test parameters. Gates proposed establishing international bodies inspired by the International Atomic Energy Agency, global civil aviation rules, and the Montreal Protocol on the ozone layer. Similar ideas for multilateral supervision were discussed at the G7 meeting in Evian in mid-June by executives including Sam Altman of OpenAI, Dario Amodei of Anthropic, and Demis Hassabis of Google.

Diplomatic outreach and security risks

Staff for Gates are currently preparing a trip to China tentatively scheduled for November, where Gates plans to meet with President Xi Jinping. Gates previously met with Xi three years ago, becoming the first foreign entrepreneur hosted by the Chinese president following the pandemic. Gates stated that China might agree to limit dangerous model releases if the United States takes the initial step. His agenda includes proposals for joint monitoring of AI systems capable of synthesizing molecules or executing biological attacks. Within the United States, President Donald Trump's administration has regulated models including Anthropic's Fable 5, while candidates in the 2026 midterm elections have supported restrictions on data center development.

These security incidents are shocking, and they should blow people's minds.