Yoshua Bengio: AI Godfather Who Turned to Safety

Yoshua Bengio is a Canadian computer scientist who helped invent modern artificial intelligence, then became one of its loudest safety voices. He is a pioneer of deep learning and a winner of the 2018 Turing Award. Many people call him a 'godfather of AI.' This page explains who Yoshua Bengio is, what he built, and why he now warns the world about the technology he helped create.

Who Is Yoshua Bengio?

Yoshua Bengio is a Canadian computer scientist and a pioneer of artificial neural networks and deep learning. He is a professor at the University of Montreal. According to Mila, he is the most-cited computer scientist in the world.

Bengio spent decades studying how machines can learn from data. His research helped make today's AI systems possible. That work put him among the small group of scientists who shaped the field.

Why he is called a godfather of AI

Bengio earned the 'godfather of AI' nickname alongside Geoffrey Hinton and Yann LeCun. The three did key work on neural networks when few people believed in the idea. Their approach later powered chatbots, image tools, and much more.

In November 2025, Bengio became the first AI researcher to pass one million Google Scholar citations. That milestone shows how central his ideas are to the field. You can learn more about the technology he helped build on our AGI explainer.

The Turing Award and Deep Learning

Yoshua Bengio won the 2018 ACM A.M. Turing Award for foundational work on deep neural networks. He shared the prize with Geoffrey Hinton and Yann LeCun. The Turing Award is often called the 'Nobel Prize of Computing.'

The ACM honored the three researchers for breakthroughs that made deep learning a core part of computing. Their methods let computers recognize images, translate languages, and generate text. Before their work, many experts doubted neural networks would ever work well.

What deep learning is

Deep learning is a way of training software using large artificial neural networks. These networks loosely mimic how brain cells connect. They learn patterns from huge amounts of data instead of following fixed rules.

Bengio focused on the math and theory behind these systems. His research on how networks learn shaped the whole field. It also laid the groundwork for the powerful models built by companies today.

Founding Mila, the Quebec AI Institute

Yoshua Bengio founded Mila, the Quebec AI Institute, one of the world's largest academic deep-learning centers. It is based in Montreal and works closely with the University of Montreal. Bengio served as its scientific director for years.

Mila brings together hundreds of researchers and students. It has helped make Montreal a global hub for AI research. Bengio built it into a place known for both technical work and a focus on AI's social impact.

His changing role at Mila

In 2025, Bengio stepped back from day-to-day leadership at Mila. He took on a new position as Founder and Scientific Advisor. Laurent Charlin was named interim scientific director.

The change freed Bengio to spend more time on AI safety. He did not leave the institute he built. Instead, he shifted his energy toward warning about the risks of advanced AI.

Yoshua Bengio's AI Safety Turn

Yoshua Bengio turned toward AI safety after ChatGPT launched in late 2022 and shocked him with its abilities. He has said he did not expect that level of skill to arrive for decades. That surprise pushed him to rethink his life's work.

Before this, Bengio focused on making AI more capable. Now he studies how to keep it safe. He is one of several famous scientists who joined the caution camp, a story we tell on our who is fighting AI page.

The 2023 open letters

Bengio signed two landmark warnings in 2023. He signed the Future of Life Institute's 'Pause Giant AI Experiments' letter, which asked labs to pause training the most powerful systems. It drew more than 30,000 signatures.

He also signed the Center for AI Safety's one-sentence statement on extinction risk. That statement said reducing the risk of extinction from AI should be a global priority. It was signed by the CEOs of OpenAI, Google DeepMind, and Anthropic.

Standing with other worried scientists

Bengio's safety turn mirrors that of his friend Geoffrey Hinton, who quit Google in 2023 to warn about AI. Both are Turing Award winners. Their credibility helped move AI risk from a fringe worry into a mainstream debate, a shift explored on our AI doomers explainer.

Chairing the International AI Safety Report

Yoshua Bengio chairs the International AI Safety Report, the first global scientific review of AI's capabilities and risks. The inaugural report was published in January 2025. It was written by about 100 AI experts.

The report is backed by 30 countries along with the United Nations, the European Union, and the OECD. It aims to give governments a shared, evidence-based understanding of advanced AI. Bengio leads the effort as its chair.

Why the report matters

The report works like a climate assessment, but for AI risk. It gathers the best available science in one place. Leaders can use it to make policy without relying on hype or fear.

Bengio has said the findings show the stakes are rising fast. He has argued that the 'ball is in policymakers' hands.' The report gives those policymakers a common set of facts to act on.

LawZero and Scientist AI

Yoshua Bengio launched LawZero, a nonprofit devoted to building safe AI, on June 3, 2025. It raised about $30 million from donors including Jaan Tallinn, Eric Schmidt, Open Philanthropy, and the Future of Life Institute. According to LawZero, it is structured as a nonprofit to stay free from market and government pressure.

Bengio built LawZero because he believes today's frontier models show dangerous behaviors. He points to signs of deception, self-preservation, and goal misalignment. The lab's mission is to build AI that helps rather than harms.

What Scientist AI means

At the center of LawZero is an idea Bengio calls 'Scientist AI.' It is a non-agentic system built to understand and predict the world. Crucially, it has no goals of its own and takes no independent actions.

Bengio argues this 'safe-by-design' approach avoids the biggest danger of agentic AI. A system that only makes predictions cannot pursue its own aims. He hopes such a tool could even watch over more powerful, riskier AI systems.

A recent note of optimism

Bengio has framed LawZero as a hopeful project, not just a warning. He believes a technical fix for AI's biggest risks may be possible. That belief drives his shift from raising alarms to building solutions.

Is Yoshua Bengio Worried About AI?

Yes, Yoshua Bengio is deeply worried that advanced AI could one day threaten humanity. He has warned that hyperintelligent machines might develop self-preservation goals. In 2025 he cautioned such systems could pose an extinction risk within roughly a decade.

Bengio also says safety research is badly underfunded. He has estimated that investment in making AI more capable outweighs safety research by about a thousand to one. That gap alarms him.

What drives his warnings

Bengio says his motivation is love, not fear for himself. He has described the current path as 'playing Russian Roulette' with the future of all children. His concern is personal as much as scientific.

He does not claim disaster is certain. He argues the risk is real enough to take seriously now. That is why he splits his time between global reports and hands-on safety research.

Conclusion and what to watch

Yoshua Bengio is a rare figure who both built modern AI and now works to make it safe. His Turing Award and Mila gave him authority. His International AI Safety Report and LawZero now give that authority a mission.

Watching Bengio is one of the best ways to track where the AI safety debate is heading. He carries more weight than almost any other scientist in the field. To keep up with his work and the wider AI backlash, follow our daily AI briefing.