“I feel lost. I feel like the world is not prepared for what's coming.”
“Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”
“I used to think it would take decades before AI could threaten us. Now I'm not sure we have that long.”
“We need international governance for AI that is at least as strong as what we have for nuclear weapons.”
“The people building these systems are not the right people to decide on their own how to govern them.”
“I am worried that the commercial pressures are overwhelming the safety considerations.”
“This technology is advancing faster than our ability to understand it, let alone control it.”
Yoshua Bengio represents perhaps the most dramatic intellectual journey in the AI field. As one of three researchers awarded the 2018 Turing Award for their work on deep learning (alongside Geoffrey Hinton and Yann LeCun), Bengio spent decades advancing the capabilities of neural networks. He has now become one of the most vocal advocates for treating advanced AI as an existential threat.
His public statements since 2023 carry the weight of a researcher who helped create the field now expressing genuine fear about where it is heading. When Bengio says he feels “lost” about the trajectory of AI development, it carries different significance than the same statement from someone who had not dedicated their career to advancing the technology.
Bengio’s shift toward alarm has been more public and emotional than Hinton’s more measured warnings. He has expressed genuine personal distress about the pace of development and the inadequacy of governance responses. His calls for international governance comparable to nuclear weapons treaties reflect a belief that the threat is of comparable magnitude.
His relationship with the commercial AI industry is complex. Bengio works closely with companies through Mila (the Quebec AI institute he founded) while also arguing that those same companies should not self-govern. He consistently pushes for external oversight by democratic institutions rather than voluntary industry commitments.
Bengio’s position is particularly influential because of his credibility as a technical pioneer. Critics of AI risk advocacy cannot easily dismiss him as uninformed about the technology. His combination of technical expertise and genuine alarm makes him one of the most compelling voices in the AI safety movement.