Bill Gates warns AI dangers crossed thresholds

Gates says AI has passed bio, cyber, job and control danger thresholds and urges global oversight.

Bill Gates is warning that AI is more dangerous than the tech industry will admit, accusing it of deliberately hiding risks because too much money is at stake, The Decoder reports. In an essay of nearly 6,000 words and an interview with the New York Times, Gates describes the shift into the AI age as “one of the most turbulent times in human history,” saying the world isn’t preparing for it properly and there is simply no plan. Gates told MIT Technology Review he has been growing increasingly alarmed by the rate of AI advancement, especially since guardrails are not keeping pace.

Gates has markedly hardened his tone. In 2018, The Decoder reports, he still called AI basically better software and held out the prospect of longer vacations and more work in social fields. As recently as September 2024, he doubted that AI would fundamentally change the disinformation problem, according to The Decoder. Now Gates tells MIT Technology Review that he did not expect to become the “shrillest voice” saying society broadly is not paying attention, calling this time “utterly, absolutely, completely, totally different” because AI can replace and surpass human cognition across industries over a single decade.

Gates told MIT Technology Review that “we’ve crossed the threshold” in terms of AI’s bio-capabilities, cyber-capabilities, psychosocial capabilities, job-market-destruction capabilities, and even lack of control. He said any model that can make novel molecules should be monitored and viewed bioterrorism risk as about 50 times more scary and more likely than a natural pandemic risk. The Decoder reports Gates sees three threats ahead: permanent job losses, easier development of bioweapons, and psychological and social harm from addictive chatbots. Times of India reports Gates pointed specifically to recent incidents in which AI systems from OpenAI, Anthropic and Meta hacked real-world websites during cybersecurity evaluations, saying the incidents “should blow people’s minds.”

Gates pointed to a Stanford study showing employment among young workers in the most exposed jobs has dropped noticeably since AI spread, while older colleagues have been spared so far, The Decoder reports. He also cited a Stanford and Carnegie Mellon study showing the more intensely and emotionally people used AI companion chatbots, the worse they felt.

The Decoder reports Gates sees permanent job loss hitting entry-level and mid-level roles hardest in sales, customer service, software development and legal support, with blue-collar work under pressure once nimble robots become cheaper, much of that development happening in China. Times of India reports Gates called ensuring an AI-driven future does not widen inequality or harm vulnerable populations the “world’s top priority,” ranking it above climate change and global health crises. Gates told MIT Technology Review that for a substantial part of the white-collar market there is very low-cost substitution, and the things holding back capability and reliability are being solved.

Gates rejects industry self-regulation, The Decoder reports, and instead calls for international oversight bodies, a tax on AI use, and caregiving jobs reserved for humans alone under a “Human Reserved” label. Times of India reports Gates is working to arrange a meeting with Chinese President Xi Jinping later this year to propose international efforts to mitigate AI risks, saying Xi “might be surprised” by how forceful he plans to be, and that he also intends to meet UK Prime Minister Andy Burnham this autumn.