Yoshua Bengio

Yoshua Bengio is the third of the three researchers awarded the 2018 Turing Award for deep learning. Working from Montréal, his lab produced foundational work on recurrent neural networks, attention mechanisms, generative adversarial networks (GANs), and — critically — word embeddings and neural language modelling, which are the direct precursors of the transformer architectures that underlie GPT, Claude, and Gemini.

Unlike Hinton, who left Google to speak freely about AI risk, Bengio has remained in academia and built his AI safety position from within the research community. He was a lead author on the 2023 "Statement on AI Risk" signed by leading AI researchers, and has been a central figure in international AI governance discussions — including at the UK AI Safety Summit at Bletchley Park in November 2023. He co-authored the 2024 paper "Managing Extreme AI Risks Amid Rapid Progress" with Hinton, Turing Award winner and Turing Institute chair Andrew Yao, and others, which called for mandatory safety evaluations and international coordination.

Bengio has explicitly framed the AI safety question in terms of democratic accountability: who decides what values AI systems encode, and through what legitimate process? This is a governance question, not just a technical one, and his framing has influenced how AI regulation discussions proceed at international bodies.