September 2026

This monthly newsletter covers recent developments and upcoming events in AI safety, ethics, and governance in Montréal.

Events

Nine of the community events below do not have a registration page yet. Dates are confirmed, times and venues are not. They will appear on the community calendar as they are posted.

Community Lunch and Ω Labs Coworking
Thursdays September 3, 10 and 24, 12:30 to 4:30 PM. Ω Labs, 3813 rue Saint-Denis, Montréal.
A recurring open afternoon at Ω Labs: lunch together, then coworking for anyone in Montréal working on AI safety, ethics, or governance. Each date has its own registration page.

LawZero presents Safety from Honesty in a Disinterested AI Predictor
Thursday, September 3, 6 to 8 PM. Ω Labs, 3813 rue Saint-Denis, Montréal.
LawZero presents the paper behind its Scientist AI: the argument that a system trained to predict truthfully, without any stake in the consequences of its own outputs, avoids acquiring implicit agency. The paper is in Research below, and reading it in advance will help.

TAISRG session
Tuesday, September 8. To be confirmed.
A session of the technical AI safety reading group, still being confirmed.

RustConf AI safety social
Thursday, September 10. Details to come.
An AI safety social on the sidelines of RustConf 2026, which runs September 8 to 11 at the Palais des congrès.

AI Incident Response Sprint
Friday, September 11 to Sunday, September 13. Details to come.
A weekend sprint on how AI incidents get detected, reported, and responded to. It lands while Ottawa’s transparency consultation, which asks exactly this question, is still open.

IVADO workshop 2: Adapting AI Governance to Sector-Specific Challenges
Monday, September 14 to Friday, September 18, 9:30 AM to 5 PM. IVADO and HEC Montréal, 950 av. Beaumont. Registration open, $70 to $500.
The second workshop of IVADO’s Governance of AI semester gives each day to one sector: media and information in democracies, artistic creation, justice and social services, and health and medicine. The Friday closing conference on cross-cutting issues is included in the registration.

AI Governance Reading Group #2
Tuesday, September 15. Details to come.
The second session of the reading group that began in August with Barnett and Thiergart on the limits of frontier-model evaluations. Reading to be announced.

ALL IN AI safety social
Wednesday, September 16. Details to come.
An AI safety social alongside ALL IN 2026, Canada’s largest AI conference, at the Palais des congrès on September 16 and 17.

Speaker series
Tuesday, September 22. Speaker and details to come.
An evening talk in the community speaker series.

Women mixer
Thursday, September 24. Details to come.
An evening for women working on or curious about AI safety, ethics, and governance in Montréal.

AI Strategy Course, inaugural session
Monday, September 28 or Wednesday, September 30, to be confirmed. Details to come.
The opening session of a five-week course on AI strategy, running weekday evenings.

Community Progress Review
Tuesday, September 29. Details to come.
An open review of what the Montréal AI safety community moved this quarter, and what it did not.

Opportunities

OBVIA: call for AI workshop facilitators, deadline September 4
OBVIA is assembling a paid roster to run awareness and discussion workshops on AI issues for the general public, community groups, unions, and student audiences. Open to students at a Québec university who live in Québec, with remuneration and travel costs covered, and an online information session on September 10 for those selected. This closes days after this newsletter goes out, so apply now or not at all.

Federal consultation on AI transparency, final call, closes September 23
ISED wants input on consistent public information about AI systems, tracking of serious AI incidents, and ways to trace what AI agents do. This is the last edition before it closes, and the questions on incident reporting are the ones this community is best placed to answer.

Policy and Governance

Four Québec parties sign the AI code of conduct for the election campaign
August 24.
The Coalition avenir Québec, the Parti libéral du Québec, the Parti québécois, and Québec solidaire have endorsed the voluntary code written by Catherine Régis (IVADO, Université de Montréal), Christian Gagné (IID, Université Laval), and Marc-Antoine Dilhac (Université de Montréal), and published by IVADO and CEIMIA. Signatories commit to clearly labelling AI-generated content, avoiding disinformation and deepfakes, protecting personal data, securing their AI systems, and training staff on AI-related electoral risks. The National Assembly had adopted a motion in March calling on parties to sign, and this is the first time they actually have, with the general election on October 5. The Parti conservateur du Québec has not signed.

Québec and Université Laval build a sovereign testing environment for AI in public services
August 21.
The ministère de la Cybersécurité et du Numérique is putting $2.1M over two years, with an option for $1M more in a third year, into a secure and sovereign test bench run with Université Laval’s Institut intelligence et données, where concrete AI uses for public services can be tried out. It lands while the Protecteur du citoyen’s investigation into AI and algorithms in Québec public services is still open, so what evaluation this bench actually requires before a system reaches citizens is worth watching.

Research from Montréal

Safety from Honesty in a Disinterested AI Predictor
June 28, revised July 10. Yoshua Bengio (LawZero, Mila, Université de Montréal) and 15 co-authors.
The formal safety case behind LawZero’s Scientist AI. Train a system to approximate Bayesian posteriors over epistemically contextualized text, and it has no stake in its own deployment, which is what the authors argue keeps implicit agency from emerging. They give conditions under which coordinated deception across queries would be statistically unlikely to survive training. LawZero presents this paper at Ω Labs on September 3.

Citizen participation in the governance of AI for urban mobility in Greater Montréal
July 14. Eddy Damaris Nono Defo, OBVIA student publication.
A study of how residents of the Montréal metropolitan region can be brought into decisions about AI systems in urban transport, and which participation mechanisms actually reach users rather than only institutional stakeholders.

Arts and Culture

“À l’heure de l’IA” continues its tour, in Longueuil October 5 to 28
Cégep de St-Félicien, September 2 to October 1. Cégep Édouard-Montpetit, Longueuil, October 5 to 28.
OBVIA’s interactive exhibition on the societal impacts of AI, eight everyday-life scenes plus modules on the ethical principles of responsible AI, left the Grande Bibliothèque at the end of August and is now touring. The October dates at Cégep Édouard-Montpetit put it back within reach of Montréal.

Have an event, opportunity, or article for next month? Send us a note at team@horizonomega.org.

This newsletter is by HΩ, researched and written with the assistance of AI.