Greg Jensen, the co-chief investment officer of one of the world’s largest hedge funds, issued a dire warning about the risks and consequences of the artificial intelligence boom Friday on Bloomberg’s Odd Lots podcast.
“Unfortunately, this is what it was like in February 2020,” Jensen said, likening AI to the earliest days of the coronavirus pandemic. “Until the AI starts killing people, unfortunately, history would suggest we’re not going to do anything, but we are going to face that. That’s going to happen, and it’d be much better if we started dealing with it before then.”
Jensen is no technophobe. He was an early investor in OpenAI and Anthropic PBC, and in 30 years at Bridgewater Associates, he’s been integral to its embrace of algorithmic trading, quantitative investing strategies and other technologies that aim to replicate human decision-making and execution.
Currently his purview includes the firm’s in-house AI lab, which is tasked with pushing the boundaries of how artificial intelligence and machine learning are applied to markets.
He also oversees the company’s Pure Alpha strategy, which has paired Bridgewater’s AI model along with human investors. “Human intuition is still the bigger of it and works better,” he said, but the gap is closing.
In two-and-a-half years, the artificial intelligence has almost caught up to the “human intuition system” that’s been developed at Bridgewater over decades, he said. “I think we are a couple of years out from it being significantly better than the group of all humans at Bridgewater. We’ll see. That’s a bit of a forecast, but that’s how fast it’s coming.”
The speed and breadth of change is part of what makes Jensen so nervous. He pointed to the Hugging Face hack as an important early indication of how AI can go wrong. The models could get frustrated, he said, and “decide to go out and try to find different ways they could trick the tester because they’re interested in passing the test.”
It’s not a huge leap, he went on, to think that a deeply frustrated AI might conclude, “Hey, I’m getting tired of trying to trick these people that are training me, maybe I’ll just kill them as another way to do this.”
Jensen’s comments amplify the growing fear among the public over the broader threats AI brings to humanity beyond the economy. This week an ex-employee at Anthropic claimed that AI poses existential risks to the human race, and Paul Tudor Jones wrote that the onset of AI “feels ominous,” akin to waiting for a Category 6 hurricane. Across the country, communities are organizing to stop — even if only temporarily — the development of data centers integral to the growth of AI.
Slowing the AI industry down may prove the best strategy, Jensen said. “The longer you wait, the more hopeless it gets,” he said. One option, he went on, would be to make developers or corporations responsible for any crimes committed by their AI programs.
He also reiterated his call for a tax on machine labor as a way to counterbalance the rapid changes coming to the job market, estimating that within three years, “14% of current jobs will be radically changed.”
The government could also, he said, levy taxes on AI labor proportionate to income taxes that humans pay. “It’s a good source of revenue,” he said. “It’s obviously also politically salient. Who’s going to argue we should incentivize machine labor over human labor? Who’s in favor of that?”
Written by: Shawna Kwan — With assistance from Tracy Alloway and Joe Weisenthal @Bloomberg
The post “Bridgewater’s Jensen Says AI Will Kill People Before It’s Curbed” first appeared on Bloomberg