Yoshua Bengio

Yoshua Bengio is a computer scientist who was born in Paris, France, in 1964 and moved to Montreal, Canada, at the age of 12. He received a bachelor's degree in computer engineering and a master's and a PhD (1991) in computer science from McGill University in Canada. After research at MIT and AT&T Bell Labs in the United States, he became a professor at the Université de Montréal in 1993. As of October 2026, he is a professor of computer science at the Université de Montréal and Co-President and Scientific Director of LawZero, a non-profit organisation that researches AI safety.

Bengio has done a great deal of research on how neural networks handle language. In a paper published in 2000, 'A Neural Probabilistic Language Model', he introduced word embeddings (a way of representing the meaning of a word as a set of numbers) into a neural language model. His group added attention (a mechanism that helps a model focus on the important parts of a sentence) to this approach and improved machine translation. With Ian Goodfellow, then his PhD student, he also developed generative adversarial networks (GANs), in which two neural networks learn by competing against each other.

Bengio founded Mila, an AI research institute in Montreal; he was its scientific director until spring 2025 and is now its scientific adviser. In June 2025 he founded LawZero to research AI that is 'safe by design'. Instead of AI that acts on its own, LawZero proposes a 'Scientist AI' that tries to understand the world, shows its reasoning and answers honestly. He also chaired the International AI Safety Report published in February 2026, and he holds a Canada CIFAR AI Chair. In 2019 he received the 2018 Turing Award with Hinton and LeCun, and in 2017 he was made an Officer of the Order of Canada. He has said that in 2023 he decided to devote the rest of his career to reducing catastrophic risks from AI and to developing AI for the benefit of all. In June 2025 he said that 'current frontier systems are already showing signs of self-preservation and deceptive behaviours'.