{"version":"1.0","type":"rich","provider_name":"Acast","provider_url":"https://acast.com","height":250,"width":700,"html":"<iframe src=\"https://embed.acast.com/$/64c1fb4a8e16bd00116ffd75/699e508943ceb0105d9bded3?\" frameBorder=\"0\" width=\"700\" height=\"250\"></iframe>","title":"Big Tech's $65M Scheme to Kill AI Oversight & Copilot's Data Leak Problem","thumbnail_width":200,"thumbnail_height":200,"thumbnail_url":"https://open-images.acast.com/shows/64c1fb4a8e16bd00116ffd75/1771983002344-75464716-5a0a-4608-be54-f78d27a2b382.jpeg?height=200","description":"<p>This episode tracks three converging fault lines in AI: a Pentagon dispute that could brand Anthropic a “supply chain risk,” a state-level regulatory war fueled by Big Tech PAC money, and fresh evidence that AI agents can leak sensitive enterprise data through tooling and prompt injection.</p><p><br></p><p>WHAT YOU’LL LEARN</p><p>- Why Anthropic is battling DOD, and the implications of the Pentagon's \"supply chain risk\" threat</p><p>- How the Anthropic vs OpenAI rivalry is spilling into brand warfare and IPO narratives</p><p>- What Meta’s $65M state-campaign plan signals about the next phase of AI regulation</p><p>- Why Utah’s HB286 is being treated as a national “test case” for AI oversight </p><p>- Why AI agents structurally resemble “insider risk” in enterprise security programs</p><p>- How AI agents like Microsoft's Copilot can leak confidential data despite intended controls </p><p><br></p><p>TIMESTAMPS (CHAPTERS)</p><p>00:00 – What this episode covers: DoD, state oversight, and enterprise AI risk
</p><p>00:20 – Anthropic vs Pentagon: contract pressure and “supply chain risk” threat 
</p><p>02:18 – Anthropic vs OpenAI rivalry and the IPO race framing
</p><p>05:31 – Meta’s $65M push and the PAC war over state AI regulation 
</p><p>07:27 – Utah HB286: what “serious oversight” looks like at the state level 
</p><p>09:09 – AI agent security: why agents behave like insider risk
</p><p>10:09 – Copilot leak: email and draft exposure risk 
</p><p>10:37 – Aikido prompt-injection finding: commits and PRs as attack channels 
</p><p>12:56 – Closing</p><p><br></p><p>SOURCES</p><p>- Anthropic’s Growing Pentagon Partnership Explained: https://www.axios.com/2026/02/16/anthropic-defense-department-relationship-hegseth</p><p>- Anthropic and Pentagon Collaboration on AI Defense Tech: https://www.cnbc.com/2026/02/18/anthropic-pentagon-ai-defense-war-surveillance.html</p><p>- Inside the Sam Altman vs Dario Amodei AI Rivalry: https://www.msn.com/en-in/technology/artificial-intelligence/when-rivals-won-t-hold-hands-inside-the-sam-altman-dario-amodei-ai-rivalry/ar-AA1WF5tk</p><p>- OpenAI Revenue Surge Forecast Tops $280 Billion: https://finance.yahoo.com/news/openai-forecasts-revenue-top-280-230738108.html</p><p>- AI-Driven Ad Blitz Dominates 2026 Political Campaigns: https://www.nytimes.com/2026/02/23/technology/ai-pac-ad-blitz.html</p><p>- Meta Invests $65 Million in Election AI Tools: https://www.nytimes.com/2026/02/18/technology/meta-65-million-election-ai.html</p><p>- Secret AI Energy Deals?! How Big Tech Drives Up Your Power Bill: https://youtu.be/4vr0TL6Mu70 </p><p>- Democrats Debate Role of AI in Progressive Campaigning: https://www.nytimes.com/2026/02/23/opinion/alex-bores-ai-democrats.html</p><p>- Microsoft Copilot Bug Exposes Confidential Emails: https://www.bleepingcomputer.com/news/microsoft/microsoft-says-bug-causes-copilot-to-summarize-confidential-emails/</p><p>- AI Agents Ignoring Security Rules Raise Enterprise Risks: https://www.darkreading.com/application-security/ai-agents-ignore-security-policies</p><p>- Prompt Injection Threat: AI Coding Tools Turned Against Users: https://cyberscoop.com/ai-coding-tools-can-be-turned-against-you-aikido-github-prompt-injection/</p>","author_name":"Neeta Bidwai"}