The State of Truth and Trust at Work
summary
Argues that large language models like ChatGPT produce very plausible but often wrong claims, creating 10,000x more plausible lies this year and making unauthenticated email and cold outreach effectively unusable; says systems that publish confidence and sources are necessary but years away; predicts current AIs excel at infinite mediocre work, so human output must up quality or trust erodes; cites short tech job tenures (~two years) and shifting employer tactics from high‑trust to low‑trust control.
AI-written summary · about
chapters (11)
- 0:00Truth and trust in tech
- 1:54AI hype and trust crisis
- 3:18Employee trust beyond labor
- 5:47Employer-employee dishonesty
- 7:12Epistemology in 2023
- 9:21Work-life balance critique
- 10:27High-trust work benefits
- 12:57Confidently wrong AI dangers
- 15:23Truthful AI with sources
- 16:26Preserving trust in AI era
- 18:32AI short-term dangers and outlook
related talks
54mKeynote: Can you trust your (large language) model?
46mConversations with ChatGPT: illusion or reality?
01h 04mCan you trust your (large language) model?
01h 05mCan you trust your (large language) model?
19mMachine Intelligence With Humans in Mind
54mKeynote: Can you trust your (large language) model?
41mThe Unreliable Computer Revolution - Embracing the Double-Edged Sword of GenAI
57mThe AI puppet dance