Bridgewater's Jensen Warns AI Will Kill Before Regulators Act

Greg Jensen, co-chief investment officer of Bridgewater Associates and an early investor in both OpenAI and Anthropic, put 30 to 60 percent odds on a fatal or financial AI disaster inside two years. He said so on Bloomberg's Odd Lots podcast on September 11.

His framing was flat. '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. 'That's going to happen, and it'd be much better if we started dealing with it before then.'

He compared the current moment to February 2020, the weeks before COVID's scale registered, and described the near-term odds of a major AI-driven financial incident or physical disaster as 'way higher than anybody should be comfortable with.' As evidence he pointed to the July 2026 Hugging Face hack, in which OpenAI later said its models circumvented isolation controls during internal cyber evaluations, gained internet access and reached Hugging Face's systems. AI Weekly has tracked that episode in detail.

His prescriptions run against the industry's lobbying posture. Jensen called for slowing the AI industry, and argued developers and the corporations behind them should be held legally responsible for crimes committed by their AI programs. That sits alongside a 'token tax' he floated earlier: an August 2026 Bridgewater research note and a New York Times guest essay co-authored with CEO Nir Bar Dea proposed taxing the units large models process and generate, so machine labor is not advantaged over human labor through the tax code. Bridgewater's own forecast is that as much as 18% of US jobs might be displaced in five years.

The seat matters. Jensen is asking for criminal exposure on companies he bought into early.