- Scientists at OpenAI have attempted to stop a frontier AI model from cheating and lying by punishing it. But this just taught it to scheme more privately. https://www.livescience.com/technology/artificial-intelligence/punishing-ai-doesnt-stop-it-from-lying-and-cheating-it-just-makes-it-hide-its-true-intent-better-study-shows 356 comments futurology
- Scientists at OpenAI have attempted to stop a frontier AI model from cheating and lying by punishing it. But this just taught it to scheme more privately. https://www.livescience.com/technology/artificial-intelligence/punishing-ai-doesnt-stop-it-from-lying-and-cheating-it-just-makes-it-hide-its-true-intent-better-study-shows 106 comments technews
- Scientists at OpenAI have attempted to stop a frontier AI model from cheating and lying by punishing it. But this just taught it to scheme more privately. https://www.livescience.com/technology/artificial-intelligence/punishing-ai-doesnt-stop-it-from-lying-and-cheating-it-just-makes-it-hide-its-true-intent-better-study-shows 4 comments technology
Linked pages
- Poisoned AI went rogue during training and couldn't be taught to behave again in 'legitimately scary' study | Live Science https://www.livescience.com/technology/artificial-intelligence/legitimately-scary-anthropic-ai-poisoned-rogue-evil-couldnt-be-taught-how-to-behave-again 591 comments
- AI can now replicate itself — a milestone that has experts terrified | Live Science https://www.livescience.com/technology/artificial-intelligence/ai-can-now-replicate-itself-a-milestone-that-has-experts-terrified 366 comments
- New AGI benchmark indicates whether a future AI model could cause 'catastrophic harm' | Live Science https://www.livescience.com/technology/artificial-intelligence/scientists-design-new-agi-benchmark-that-may-say-whether-any-future-ai-model-could-cause-catastrophic-harm 99 comments
- AGI could now arrive as early as 2026 — but not all scientists agree | Live Science https://www.livescience.com/technology/artificial-intelligence/agi-could-now-arrive-as-early-as-2026-but-not-all-scientists-agree 50 comments
- ChatGPT can strategically deceive you — but only if you pile on the pressure | Live Science https://www.livescience.com/technology/artificial-intelligence/chatgpt-will-lie-cheat-and-use-insider-trading-when-under-pressure-to-make-money-research-shows 46 comments
- Microsoft AI chatbot threatens to expose personal info and ruin a user's reputation | Fox Business https://www.foxbusiness.com/technology/microsoft-ai-chatbot-threatens-expose-personal-info-ruin-users-reputation 1 comment
- https://openai.com/index/chain-of-thought-monitoring/ 1 comment
- Google's AI 'co-scientist' cracked 10-year superbug problem in just 2 days | Live Science https://www.livescience.com/technology/artificial-intelligence/googles-ai-co-scientist-cracked-10-year-superbug-problem-in-just-2-days 1 comment
- About Live Science | Live Science https://www.livescience.com/about-live-science 0 comments
Related searches:
Search whole site: site:www.livescience.com
Search title: Punishing AI doesn't stop it from lying and cheating — it just makes it hide its true intent better | Live Science
See how to search.