Risks of AI Risk Policy: Five Lessons
summary
AI policy proliferation hits specifics: 21 draft standards appeared last year and two flagships are NIST AI RMF and the draft EU AI Act; real systems contain hundreds of ML models while ML code is often ~5% of production code; known attacks include a PyPI supply‑chain trojan against a torch package and model inversion that extracts private training data via API queries; human studies show 67% poured juice into a plant and 40/40 unlocked laptops for a robot, and 70% of law experts trust misleading explanations.
AI-written summary · about
chapters (7)
- 0:00AI risk policy intro
- 4:00Automation bias in AI
- 10:42Defining AI systems
- 16:24AI standards vagueness
- 25:34Who shapes risk frameworks
- 30:29Consumer trust in AI
- 32:34Call to ethical AI
related talks
53mAI: Law, Policy and Common Sense Suggestions on How to Stay Out of Trouble
01h 00mAI: Law, Policy, and Common Sense Suggestions to Stay Out of Trouble
36mResponsible AI: In Theory, Practice, and Policies
31mPanel: Navigating the AI Policy Landscape
45mHardening AI/ML Systems - The Next Frontier of Cybersecurity
01h 05mNavigating the Security and Privacy Landscape of Modern AI
24mTechnical Tracks
36mAvoiding 5 missteps that undermine your AI readiness and success