Being Good: An Introduction to Robo- and Machine Ethics
summary
Three concrete failure modes decide whether to watch: a medical irradiator (Therac-25) replaces hardware interlocks with assembly code and concurrent bugs and causes about six deaths; an auto maker’s defeat device hides emissions and links to roughly 59 premature deaths; a 2016 DAO smart contract loses four million ether via a reentrancy bug (about $70M then, ~$840M later). Also: facial recognition and predictive policing models learn bias (a decision tree reaches ~77% accuracy by splitting on race), and autonomous cars expose unclear liability.
AI-written summary · about
chapters (19)
- 0:00Intro
- 3:15Agenda
- 7:47What is being good
- 10:22The Rac25
- 12:55What happened
- 14:07Standard of Care
- 16:11Test Mode
- 20:25Ethereum Smart Contracts
- 21:54Ethereum Hard Fork
- 23:08Machine Ethics
- 24:41Apple Face ID
- 26:12Password hashes
- 26:49Moral weighting
- 30:07Bias data
- 30:35Weinsteins conjecture
- 31:27Selfdriving cars
- 32:26The trolley problem
- 34:35Teaching ethics
- 38:40Code reviews
related talks
39mCan We Still Build Technology for the Right Reasons
47mThe Ethics of Software - some practical considerations
41mMachine Ethics and Emerging Technologies
16mOn the ethical challenges raised by robots powered by AI
16mFrom Existential to Existing Risks of Generative AI: A Taxonomy of Who Is at Risk,...
43mEthical Engineering
15mWhen Machines See: Navigating Privacy
57mToxic Trust: The Psychology of Automation Bias in Generative AI