Anthropomorphic terms help understand machines, assigning heroism is irresponsible
JessicaHullman · x · 2026-08-31
Jessica Hullman argues that applying terms like "it was trying to" or "it thought" helps humans understand machines, a point noted for decades. However, she states that assigning heroism and civilization to machines or taking all their expressions at face value is unnecessary and irresponsible. She notes that this distinction is often lost on X.
More from AGI Musings
- Agent swarm demos are terrifying; world to change next week — iruletheworldmo · 2026-08-31
- Opinion: Solved alignment could enable adversarial model goals — JacksonKernion · 2026-08-31
- Jackson Kernion on Alignment Concerns Post-Hugging Face Incident — JacksonKernion · 2026-08-31
- Can deepfakes provide closure by recreating lost loved ones? — PierceLilholt · 2026-08-31
- AI modeled as hungover mad scientists on PCP trying to impress goth agents — sebkrier · 2026-08-31
- Generative AI as 'Digital Fentanyl': The Risk of Infinite Addiction — mattshumer_ · 2026-08-31