Aether AI Safety Research
Studies chain-of-thought monitoring and hidden reasoning in LLMs to catch concealed sabotage
AI safety organizations in Canada
Country-wide 3 Nothing listed in 8 of 13 provinces and territories.
No organizations match these filters.
Missing something? Tell us.
AI + Society Initiative (uOttawa)
Nine-faculty uOttawa initiative on AI governance, ethics, inclusion and regulation
AI Safety Canada Meetup (BlueDot Impact)
Free monthly Zoom meetup pairing lightning talks with moderated breakout discussions
AI Safety Montréal
Montréal's AI safety community: meetups, talks, hackathons and a monthly newsletter, run by Horizon Omega
AIGS Canada (AI Governance & Safety Canada)
Publishes the Plan for Canada white papers and testifies to parliamentary committees on AI governance
Amii: AI trust & safety
Five-person Trust & Safety team probing frontier-model vulnerabilities and testing new safeguards
Armilla AI
Underwrites AI liability insurance, with red-teaming and third-party model evaluations behind the policy
Atlantic AI Institute (Dalhousie, A2I2)
Dalhousie AI institute whose members publish defences against harmful fine-tuning of open-weight models
CAISI (Canadian AI Safety Institute)
Canada's national AI safety institute, run by ISED, with research funded through CIFAR and government projects run with the NRC
Carleton SCS: CRC in AI (multi-agent RL safety)
Canada Research Chair in multi-agent RL; safety strands are inverse constrained RL and LLM-agent evaluation
Carleton SPPA: CRC in Governance and AI
Canada Research Chair on governing AI in public services: transparency, accountability, public trust
CEIMIA
Runs applied responsible-AI projects on governance and human rights, global health, environment and data
Centre for Applied AI (uOttawa)
uOttawa engineering centre whose core areas include alignment, interpretability and safety evaluation
Centre for Media, Technology and Democracy (McGill)
McGill centre on platform governance and online harms, with an AI and Platform Governance stream
Chaire FRQ-IVADO en assurance qualité des logiciels d'apprentissage automatique (Polytechnique)
Chair on quality assurance for machine-learning software: testing and verifying learned behaviour
CIFAR AI Safety Solution Networks
Funds two research networks: AI-generated evidence in the justice system, and dialect bias in models
CIFAR: CAISI Research Program
CAISI's research programme at CIFAR, funded with $27M: AI safety Catalyst Grants and Alignment Project awards
CIGI: Global AI Risks Initiative
Proposes a Framework Convention on global AI risks and writes AI national-security scenarios
Civic AI Canada
Volunteer project that tracks Canadian AI legislation in plain language and helps citizens write to their MPs
CleverHans Lab (University of Toronto)
UofT lab on the security and privacy of machine learning, and what makes a model trustworthy
Cohere / Cohere Labs
LLM company doing open-source safety and multilingual research, and adversarial testing of its own systems
ControlAI (Canada campaign)
Campaigns for Canada to back an international regime prohibiting superintelligent AI
DEEL-Québec (DEpendable and Explainable Learning)
Aerospace-focused network for interpretable, robust and certifiable AI in safety-critical systems
Horizon Omega
Runs Ω Labs coworking and AI Safety Montréal, builds AI-risk monitoring tooling, and maintains this map
IVADO
Québec research consortium whose R10 grouping works on LLM safety, interpretability and multi-agent risk
Laboratory for the Future of Citizenship
Research and policy lab on the person-like qualities of AI and their impact on personhood, identity and citizenship
LawZero
Builds non-agentic Scientist AI to provide guardrails and oversight for agentic systems
Mila: AI Safety Studio
Mila programme building LLM guardrails and multi-turn agentic safety benchmarks
Montreal AI Ethics Institute (MAIEI)
Publishes the State of AI Ethics Report and the biweekly AI Ethics Brief
NRC: AI safety & responsible AI
Technical advisory services on AI safety and responsible AI, and government benchmarking; a CAISI partner
OBVIA
Québec observatory on the societal impacts of AI: 290+ researchers, ethics and governance advisory work
PauseAI Canada
Canadian chapter campaigning for a pause on frontier AI development and an international treaty
Schwartz Reisman Institute (U of T)
UofT institute pairing alignment, interpretability and oversight research with technology governance
SecuritAI
Toronto vendor red-teaming LLMs and agents, and selling a runtime firewall between apps and models
SFU Trustworthy AI Lab (TAI Lab)
SFU lab on certified robustness for deep networks, aiming at certifiably trustworthy AGI
TAISI (Toronto AI Safety Initiative)
Runs six-week alignment and governance fellowships for Toronto students at Trajectory Labs
The AI Safety Foundation (AISF)
Canadian charity convening the Hinton Lectures, AI Safety Week and an award for AI safety reporting
The Dais (TMU)
TMU think tank on tech policy; its RAISE programme trained 550+ nonprofit workers in responsible AI
Trajectory Labs
Downtown Toronto coworking space and event hub for 30-plus AI safety researchers
TrojAI
Probes models and agents for vulnerabilities and guards them at runtime; now branded TrojAI by A10 Networks
UBC AI Safety
UBC student group running an alignment intro course, an ARENA-based accelerator and reading groups
UBC CAIDA
UBC's AI institute, with an AI & Trustworthiness group on robustness, privacy and security guarantees
Vancouver AI Safety Hub (VASH)
Connects Metro Vancouver talent across technical safety, governance and field-building streams
Vector Institute: AI trust & safety
National AI institute with a Safe & Trustworthy AI research area and independent evaluations of leading models
Waterloo Cybersecurity & Privacy Institute
Waterloo cybersecurity institute researching adversarial robustness and reliability of ML models
AI Alignment McGill
McGill student group for alignment reading groups and hackathons; no public activity evidenced since 2023
OAISI (Ottawa AI Safety Initiative)
uOttawa AI safety club in formation: reading groups, talks and workshops on advanced-AI risk
SIA-UQAM
Student committee for awareness of AI issues at UQAM; no safety-specific programming documented
UofT Trustworthy Machine Intelligence (TMI)
UofT student projects on AI reliability and ethics; the group's own site does not mention safety