Follow
- X@Yoshua_Bengio
- Bluesky@yoshuabengio.bsky.social
- LinkedIn@yoshuabengio
- Websiteyoshuabengio.org
- Wikipediaen.wikipedia.org
Yoshua Bengio
Turing Award-winning deep-learning pioneer who chairs the International AI Safety Report and co-chairs the UN scientific panel on AI, warning that frontier systems show deception and self-preservation and that science cannot yet rule out catastrophic harm.
Roles
- 1993 – presentFull Professor, Department of Computer Science and Operations Research, Université de MontréalSource →
- 1993 – Mar 2025Founder and Scientific Director, Mila – Quebec AI InstituteFounded as the LISA lab in 1993; stepped down as Scientific Director on 28 March 2025 to focus on AI safety.Source →
- Nov 2023 – presentChair, International AI Safety ReportMandated by the Bletchley Park summit; full reports published January 2025 and February 2026.Source →
- Mar 2025 – presentFounder and Scientific Advisor, Mila – Quebec AI InstituteSource →
- Jun 2025 – presentCo-President and Scientific Director, LawZeroNon-profit launched 3 June 2025 to build non-agentic "Scientist AI" as a safety guardrail.Source →
- Mar 2026 – presentCo-Chair, UN Independent International Scientific Panel on AIElected alongside Maria Ressa at the panel's inaugural meeting.Source →
On the record
23 Sept 2026 · Testimony · UN Security Council
The dangers are real and imminent.
Co-Chair, UN Independent International Scientific Panel on AI at the time
Said after urging the Council not to dismiss the AI companies' own warnings as a marketing stunt, on the ground that the behaviour is well documented and validated by independent experts. Sourced to The Canadian Press copy, which prints it as a complete sentence; Al Jazeera and The Next Web both carry the same words and close them on an attribution comma. CTV News, the canonical of the same Canadian Press report, was fetched and carries none of the sentences.
AI’s dangers ‘real and imminent,’ Canadian AI pioneer Yoshua Bengio tells UN →
23 Sept 2026 · Testimony · UN Security Council
This Council faces an unprecedented threat … one that none of its members would choose, that none can contain alone and that does not respect the borders we defend.
Co-Chair, UN Independent International Scientific Panel on AI at the time
Opening the Security Council's first meeting on the safety risks of increasingly capable AI, convened by France, where he briefed the Council as co-chair of the UN Independent International Scientific Panel on AI alongside the chief executives of OpenAI, Anthropic and Hugging Face. The two halves cross the UN's attribution tag ("warned Yoshua Bengio, co-Chair of the UN Independent International Scientific Panel on AI — ") and are joined with an ellipsis; the final period is inside the closing quotation mark on the page.
21 Sept 2026 · Statement · Independent International Scientific Panel on AI
Researchers have long warned that three conditions could lead to loss of control: a misaligned goal, the capability to pursue it, and an environment that allows it. This summer, all three came together in a real system, not a laboratory. Since this is not an isolated observation of misaligned goals, this raises serious questions about the way AI agents are currently trained.
Co-Chair, UN Independent International Scientific Panel on AI at the time
Issued as co-chair on the release of the Panel's first thematic brief, an assessment of the breach of Hugging Face's systems by AI agents under evaluation at OpenAI between May and July 2026. The Panel's own page for the brief carries the summary but not the quotation, so the source is the UN Economic Commission for Africa's copy of the Panel's release, which carries the statement whole and dates it 21 September 2026.
1 Jul 2026 · Statement · UN News
With growing evidence of deceptive AI behaviour, science currently cannot guarantee that as capabilities continue to increase, AI will not cause catastrophic harm, either on its own or due to malicious users.
Co-Chair, UN Independent International Scientific Panel on AI at the time
Speaking as co-chair of the UN Independent International Scientific Panel on AI at the release of its first assessment.
‘The science is here’: UN chief welcomes first global AI assessment →
26 Aug 2025 · Interview · TIME
I'm the most-cited computer scientist in the world. You'd think that people would heed my warnings.
Co-President and Scientific Director, LawZero at the time
3 Jun 2025 · Post · yoshuabengio.org
the risk of losing control is all too real, but competition between companies and countries incentivizes them to accelerate without sufficient caution
Co-President and Scientific Director, LawZero at the time
Launch essay for LawZero, citing frontier-model deception and self-preservation behaviour.
30 May 2023 · Statement · Center for AI Safety
Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.
Founder and Scientific Director, Mila – Quebec AI Institute at the time
Listed as a signatory as Professor of Computer Science, U. Montreal / Mila.
22 May 2023 · Post · yoshuabengio.org
A potentially rogue AI is an autonomous AI system that could behave in ways that would be catastrophically harmful to a large fraction of humans, potentially endangering our societies and even our species or the biosphere.
Founder and Scientific Director, Mila – Quebec AI Institute at the time
Blog essay written two months after he signed the FLI pause letter.