In the last few days Anthropic have released an impressive honest account of how all models blackmail, no matter what goal they have, and despite prompt warnings, and other preventions. But do these models *want* this?

Thanks to Storyblocks for sponsoring this video! Download unlimited stock media at one set price with Storyblocks: storyblocks.com/AIExplained


AI Insiders ($9!): https://www.patreon.com/AIExplained

Chapters:
00:00 - Introduction
01:20 - What prompts blackmail?
02:44 - Blackmail walkthrough 
06:04 - ‘American interests’
08:00 - Inherent desire?
10:45 - Switching Goals
11:35 - Murder
12:22 - Realizing it’s a scenario? 
15:02 - Prompt engineering fix?
16:27 - Any fixes?
17:45 - Chekov’s Gun
19:25 - Job implications
21:19 - Bonus Details

Report: https://www.anthropic.com/research/agentic-misalignment
30 Page Appendices: https://assets.anthropic.com/m/6d46dac66e1a132a/original/Agentic_Misalignment_Appendix.pdf
Announcement: https://x.com/AnthropicAI/status/1936144602446082431?ref_src=twsrc%5Egoogle%7Ctwcamp%5Eserp%7Ctwgr%5Etweet
OpenAI Files: https://www.openaifiles.org/
Grok 4 News: https://x.com/RonFilipkowski/status/1936372579607912473
Claude 4 Report Card: https://www-cdn.anthropic.com/6be99a52cb68eb70eb9572b4cafad13df32ed995.pdf
New Apollo Research: https://www.apolloresearch.ai/blog/more-capable-models-are-better-at-in-context-scheming
Interesting Reflections: https://nostalgebraist.tumblr.com/post/785766737747574784/the-void


Non-hype Newsletter: https://signaltonoise.beehiiv.com/