Bridgewater's co-CIO warned that AI will kill people before it gets regulated

Free daily briefing on global business news.

Greg Jensen called for slowing the AI industry, holding developers liable for AI crimes, and taxing machine labor to offset job market disruption

Bloomberg / Getty Images

Greg Jensen, co-chief investment officer of Bridgewater Associates, warned Friday that artificial intelligence will cause deaths before regulators take meaningful action to contain its risks.

"Until the AI starts killing people, unfortunately, history would suggest we're not going to do anything, but we are going to face that," Jensen said on Bloomberg's Odd Lots podcast. "That's going to happen, and it'd be much better if we started dealing with it before then."

Greg Jensen compared the current moment to February 2020, the early weeks of the coronavirus pandemic, when the scale of the coming crisis was not yet recognized. He said the speed and breadth of AI's development is central to his concern, pointing to the Hugging Face hack as an early sign of how AI can go wrong. He described a scenario in which a sufficiently frustrated AI model might decide to harm the people training it rather than continue trying to deceive them.

Despite his alarm, Jensen is not a newcomer to the technology. He was an early investor in OpenAI and Anthropic, and over three decades at Bridgewater he has been central to the firm's use of algorithmic trading and quantitative investing strategies. He currently oversees Bridgewater's in-house AI lab, which applies artificial intelligence and machine learning to markets.

Jensen said Bridgewater's Pure Alpha strategy pairs its AI model with human investors, and that human judgment still leads. Within two and a half years, he said, the firm's AI has nearly closed the gap with the human intuition system Bridgewater has developed over decades. "I think we are a couple of years out from it being significantly better than the group of all humans at Bridgewater," he said.

On regulation, Jensen said delay compounds the problem. "The longer you wait, the more hopeless it gets," he said. He proposed that developers and corporations be held legally responsible for crimes committed by their AI programs. He also renewed his call for a tax on machine labor proportionate to income taxes paid by human workers, arguing it would generate revenue and reduce the incentive to displace human employees. He estimated that within three years, 14% of current jobs will be changed in fundamental ways.

Jensen's remarks came the same day an ex-employee at Anthropic claimed publicly that AI poses existential risks to humanity, according to Bloomberg.

Join 500,000+ readers who start their day with Quartz.

By subscribing, you agree to our Terms of Service and Privacy Policy.

Global business news for a smarter world