Moral Outsourcing: Humanity in the age of AI
Talkat Open Source Leadership Summit 2019 · on Tuesday, 2019-03-12· at 16:20· The Ritz-Carlton Ballroom·
summary
Argues that bias and harm come from human choices not magic and cites 85% of online interactions managed by AI by 2020 and a $100B industry by 2025, shows Fairness Tool (an Accenture algorithmic-bias detector) deployed at clients, calls out COMPAS/Northpointe and predictive policing as concrete failures that amplify racist data, notes OpenAI’s withheld language model as a governance misstep, and proposes community-led governance, a released governance toolkit and institutionalized constructive dissent to fix accountability.
AI-written summary · about
chapters (10)
- 0:00Start
- 4:51Banality of Evil
- 7:46Why Predictive Policing Is Problematic
- 9:24Moral Outsourcing
- 9:46The Problem with Moral Outsourcing
- 12:50How Do We Create Good Governance of Ai Systems
- 13:05Community Led Governance
- 13:29Constructive Dissent
- 17:37Create an Ethical Culture
- 18:08Complying with Licenses
related talks
05mToward Ethical AI
24m2025: year of Ai dilemma - ethics, regulations and innovations
29mResponsible AI Challenges and Recommendations
56mAn AI with an Agenda: How Our Cognitive Biases Leak Into Machine Learning
28mEthical AI?!
38mSocial Implications of Bias in Machine Learning
35mFrom hype to responsibility: what works & matters for whom in data and AI?
26mBuilding ethical AI: Transparency, bias mitigation, and trust