Cohen, speaking at the Silvia Investor Summit alongside entrepreneur and investor Anthony Pompliano, delivered a stark warning about the trajectory of artificial intelligence and robotics.
The GameStop Corp. (NYSE: GME) CEO said he does not expect machines to attack people first, placing the blame for any future conflict squarely on human frustration.
His concern centers on humans becoming “so fed up by how productive and how much smarter the robots are than us” that they begin attacking machines, which then retaliate.
That chain of events, Cohen warned, is how it “ends up being man versus machine,” adding, “whether that ends up being World War III, I don’t know.”
Beyond physical conflict, Cohen expressed deeper concern about the economic consequences of machines rapidly outpacing human productivity across both white-collar and manual labor sectors.
He said the pace of technological change leaves humans with “very little utility,” a condition he argued would drive unemployment higher and suppress consumer spending.
High unemployment would then create “a domino effect that is not good,” Cohen said, framing the economic risk as potentially more immediate than any armed conflict between humans and machines.
For desk jobs, Cohen stated plainly that “the machine is far superior,” and warned that robots are “probably less than a few years away” from surpassing humans in physical work as well.
Cohen was not alone in raising concerns at the intersection of AI and societal risk, with JPMorgan Chase and Co. CEO Jamie Dimon comparing AI’s impact to the Industrial Revolution but warning it is moving “more like a tsunami.”
Dimon raised specific concerns about AI agents authorized to make payments, cautioning that customers would need to share their bank codes without knowing if those agents “are going to act properly on your behalf and they may not.”
Demis Hassabis, chair of Alphabet Inc.’s Google DeepMind, described attacks involving AI agents as “early warning shots” that could escalate into major incidents within the next couple of years without stronger security measures in place.
Adding further weight to the debate, a Lancet Commission assessment identified malicious use of AI and nuclear conflict as the two threats carrying extinction-potential by 2100, while stressing that the likelihood of either scenario remains low.
The convergence of warnings from business leaders, technologists, and public health researchers signals that concerns about AI’s destabilizing potential are no longer confined to academic circles or science fiction.
