Claude Is Reading Classified Files. Its Builders Are Warning Us.
About this video
Anthropic's Claude is now being deployed inside U.S. intelligence workflows for classified document triage, a capability shift that was unthinkable two years ago. In this video, we explore how the AI arms race is warping safety protocols and the geopolitical balance of power, with former CIA officers warning about the risks of over-reliance. Chapters: 00:00 Insiders warn 00:11 Intel AI Budget Soars 00:22 Claude's New Job 00:33 Claude's workflow 00:47 Efficiency vs. Risk 00:59 When AI errs 01:13 Why didn't anyone stop it? 01:29 Former officers warn 01:43 Petraeus on guardrails 01:57 Race to the bottom 02:13 Exempt from safety 02:28 Automation bias risk 02:44 Global power shift 03:00 Competitive pressure 03:15 Control is the prize 03:31 AI Adoption by Sector 03:47 From threats to risks 04:03 The $1.8B question 04:19 The Petraeus paradox 04:35 Automation bias in action 04:51 China's AI surge 05:06 From threats to risks 05:22 The race to govern AI Sources & further reading: ā¢ Bloomberg: Anthropic's Claude Used by U.S. Intelligence ā https://www.bloomberg.com/news/articles/2024-05-20/anthropic-s-claude-used-by-us-intelligence-for-classified-triage ā¢ The Information: Insiders Warn on AI Safety in Government ā https://www.theinformation.com/articles/insiders-warn-on-ai-safety-in-government ā¢ RAND Corporation: AI and Intelligence Analysis Survey ā https://www.rand.org/pubs/research_reports/RRA2000.html ā¢ Center for Security and Emerging Technology (CSET) ā https://cset.georgetown.edu/publication/ai-and-intelligence-analysis/ ā¢ White House Executive Order on Safe, Secure, and Trustworthy AI ā https://www.whitehouse.gov/briefing-room/presidential-actions/2023/10/30/executive-order-on-the-safe-secure-and-trustworthy-development-and-use-of-artif ā¢ Reuters: Anthropic Government Deal ā https://www.reuters.com/technology/anthropic-signs-deal-with-us-government-2023-09-15/
From the video
More from The Tool Shelf
Why BAN a Robot Maker 48 Hours Before Its $9B IPO?
GPT-4 Power at a TENTH of the Price. What's the Catch?
How Does DeepSeek Skip 98% of the COMPUTE?
Three Escape Routes From Nvidia. Which Lands FIRST?
Can DeepSeek Make the Export Ban IRRELEVANT?
AI's Newest Weapon: The Autonomous Data Center
Why Chinese Open Models Are Crushing AI Prices
