Panel Discussion: Ethical AI - From talk to practice
summary
Panelists move from rhetoric to concrete practice: building an ethics practice at Whisper that must balance crisis alerts, one‑time passwords and marketing aimed to send fewer messages; a student project shows a counterfactual fairness experiment works on synthetic causal data but breaks on real data when the causal model is wrong; an outright refusal of an autonomous‑weapon/feral‑animal project; recommendation to use the Equitas auditing tool and Rachel Thomas’s practical data ethics course as immediate, actionable resources.
AI-written summary · about
chapters (13)
- 0:00Intro
- 0:25Welcome
- 2:02Introductions
- 3:03Annas background
- 4:53Lizzy background
- 6:22Fiona background
- 9:25Policies and frameworks
- 14:35Who is in the room
- 17:05Challenges
- 19:17Cards of Ethics
- 20:21Blackbox AI
- 21:57Turning down projects
- 24:23Recommendations
related talks
56mEthics in Machine Learning Panel
44mPanel: Building responsible, human-centric AI models
30mFireside chat: Breaking the wall of unethical AI
31mEthics of algorithms - from theoretical discussion to practical application
33mEthics in Machine Learning Panel Q&A
47mEthics in AI Panel
36mKeynote: Ethics in AI: Building Fair and Just Intelligence Systems
45mEthical use of AI: seeking, understanding, and countering the risks