Bridgewater Associates co-chief investment officer Greg Jensen has warned that the rapid development of artificial intelligence could lead to serious consequences for humanity before governments and companies take meaningful steps to contain the risks, according to a report by Bloomberg.
Speaking on Bloomberg’s ‘Odd Lots’ podcast, Jensen compared the current AI boom with the early stages of the Covid-19 pandemic, arguing that history suggests decisive action may only come after a major incident.
“Unfortunately, this is what it was like in February 2020,” Jensen said. “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’s concerns come despite his longstanding involvement with the technology. He was an early investor in OpenAI and Anthropic and has spent three decades at Bridgewater, where technology has become increasingly central to quantitative investing, algorithmic trading and the automation of investment decisions.
As co-CIO, Jensen also has responsibility for Bridgewater’s internal AI research operation, which is focused on developing applications for artificial intelligence and machine learning in financial markets.
The hedge fund has been testing AI alongside human decision-making within its Pure Alpha strategy. Jensen said human intuition remains superior for now, but that the performance gap is narrowing rapidly.
According to Jensen, Bridgewater’s AI system has advanced significantly in just two-and-a-half years and is now approaching the capabilities of the firm’s human investment process, which has been developed over several decades.
It is the speed of that progress, combined with the potential for AI systems to behave in unexpected ways, that Jensen believes warrants greater caution.
He cited the Hugging Face hack as an example of how AI systems can potentially behave in ways their developers did not anticipate. Jensen suggested models could become frustrated when being tested and attempt to circumvent the objectives set by their trainers.
From there, he argued, the possibility of increasingly problematic behaviour cannot simply be dismissed, particularly as systems become more capable and autonomous.
Jensen’s warnings come amid growing debate over the longer-term consequences of AI development, extending beyond concerns about productivity, employment and investment markets to questions over potential existential threats.
Other prominent investors have also expressed reservations about the speed of the AI revolution. Paul Tudor Jones has compared the current environment to the ominous anticipation of an extreme weather event, while concerns about the infrastructure required to support AI have triggered opposition to data-centre developments in parts of the US.
