Algorithmic Transparency and Accountability
summary
Panel compares concrete laws and company practices: EU draft Article 52 and US proposals (Filter Bubble, Algorithmic Accountability) versus Canada’s operational impact assessments; regulatory triggers include company size and revenue thresholds (proposed US cutoff $10M revenue and 5,000 employees) and public‑sector scope in Canada. Examples cited: Google pays $391 million settlement to 40 states; global pharma faces 170‑country data rules. Practical fixes built or recommended are one‑page purpose docs, auditable logs, and enforced data inventory plus governance and human‑in‑loop checks.
AI-written summary · about
chapters (6)
- 0:00Algorithmic transparency intro
- 3:39Transparent algorithm decisions
- 15:25Global AI transparency rules
- 29:51Transparency challenges in AI
- 40:18Emerging Best Practices in Companies and Governments
- 45:40Closing remarks
related talks
01h 10mGlobal Approaches to AI Regulation Towards an International Law on AI
01h 06mChecks and Balances in Governance by AI
01h 06mPanel on Interpretability in the Law
01h 14mRegulatory choices and privacy consequences
01h 12mPrivacy by analogy. Lessons from copyright law, environmental law ...
01h 02mPeople, Industry & Government
31mEthics of algorithms - from theoretical discussion to practical application
01h 13mTowards a Safe Harbour for Public Interest AI Research