Margaret Mitchell: AI values eras moved from fairness to safety, security next in 2027
mmitchell_ai · x · 2026-10-29
AI researcher Margaret Mitchell sketches the eras of AI-and-values discourse: 2019 was the era of fairness and bias, 2023 shifted to safety, and she predicts 2027 may become the era of security — tracking the field's evolution from algorithmic bias to model safety to security incidents.
More from AGI Musings
- Prada hires ex-Disney legend John Pomery, spotlighting hand-drawn art in AI era — ZeroStateReflex · 2026-10-29
- Dead internet theory is false — but dead ragebait theory is true, argues moultano — moultano · 2026-10-29
- Ask HN: What Do You Believe or Not Believe About AI? — BenBajarin · 2026-10-29
- AI will reward systems thinkers: Sinead Bovell on expertise beyond professional boundaries — avicgoldfarb · 2026-10-29
- STEM expertise is becoming an infohazard, researcher warns amid ETH-result shock — coherence · 2026-10-29
- flowersslop: AI gives the poor free expert-level help, yet critics want to burn it down — flowersslop · 2026-10-29