Should AI assistants show how uncertain they are? A design debate
Icy_Pack_8733 · reddit · 2026-09-28
- A Reddit discussion asks whether AI assistants should surface uncertainty more visibly, since confident answers on incomplete or wrong info can be harmful in health, financial, legal, or educational contexts.
- Ideas include separating verified facts from guesses and flagging what needs checking.
- Concern: percentage confidence scores may create a false sense of precision. Thread weighs source links, explicit "I'm not sure," and multi-answer explanations.
More from AGI Musings
- The next winning companies will be built around syncing human and agent time — ilparone · 2026-09-28
- WIRED: 22% of Companies Already Put AI Agents on Org Charts, and No One's Ready — nordicinst · 2026-09-28
- Longevity Researcher: Work Is About to Become Optional for Many People — rand_longevity · 2026-09-28
- Sarcastic Take: Models Get Kinder as They Scale, Safety Work 'Misaligns' Them — repligate · 2026-09-28
- Guardian: AI Leaders Knew the Extinction Threat Decades Ago but Pressed On Anyway — nordicinst · 2026-09-28
- Justin Skycak: You Haven't Earned the Right to Automate Until You've Felt the Manual Friction — repligate · 2026-09-28