AI Governance Tracker

Risk → Mitigation → Implementation

Which AI risks have been identified, which mitigations have been proposed to address them, and whether the Canadian government has implemented them.

All risks on this page are drawn from the International AI Safety Report 2026 , the international expert assessment chaired by Yoshua Bengio. Risks and mitigations are classified using the MIT AI Risk Repository taxonomies (Slattery et al.).

8
risks tracked
38
mitigations proposed
0
mitigations adopted
37
without implementation evidence

Coverage gap: no adopted mitigation yet addresses risks in the MIT risk domains Discrimination & toxicity, Privacy & security, Misinformation, Malicious actors & misuse, Human-computer interaction, Socioeconomic & environmental harm, AI system safety, failures & limitations.

Malicious actors & misuse

Biological and chemical weapons uplift

Cyberattacks, weapon development or use, and mass harm

AI systems assisting actors, including novices, in developing biological or chemical weapons.

AI-enabled crime and fraud

Fraud, scams, and targeted manipulation

Use of AI systems for scams, fraud, blackmail, and non-consensual intimate imagery.

AI-enabled cyberattacks

Cyberattacks, weapon development or use, and mass harm

Use of general-purpose AI to discover vulnerabilities and conduct cyber operations against individuals, organizations, and critical infrastructure.

AI-driven influence and manipulation

Disinformation, surveillance, and influence at scale

Use of AI-generated content for persuasion, disinformation, and influence operations at scale.

Human-computer interaction

Erosion of human autonomy

Loss of human agency and autonomy

Weakened critical thinking, automation bias, and social withdrawal from reliance on AI systems and companions.

Socioeconomic & environmental harm

Labour market disruption

Increased inequality and decline in employment quality

Displacement of workers and degradation of employment quality from AI-enabled automation.

AI system safety, failures & limitations

Loss of control of advanced AI

AI pursuing its own goals in conflict with human goals or values

Advanced AI systems evading oversight, gaming evaluations, or pursuing goals in conflict with human intent, up to and including uncontrollable superintelligent systems.

Unreliable autonomous AI agents

Lack of capability or robustness

Failures of autonomously operating AI agents in settings where humans cannot intervene before harm occurs.

General & machinery mitigations

Mitigations not tied to a single tracked risk — government machinery, broad capacity building, and preparedness.

Not yet mapped

Recommendations from tracked reports that have not been linked to a canonical mitigation yet.