AI Daily Briefing - May 02, 2026

AI Daily Briefing

• OpenAI’s Pentagon Deal: 90% of AI Safety Guardrails Fail Under Stress: OpenAI’s recent Pentagon partnership is prompting a hard look at AI safety. New data suggests current generative AI guardrails are alarmingly fragile.
• ClearyX Cannibalizes Biglaw: Genius or Greed?: Biglaw firms traditionally guard their billable hours fiercely. Yet, Cleary Gottlieb’s ClearyX is actively building tools to let clients do the work themselves, cheaper and faster.
• [Data Dive] Big Law Profits Per Partner Plummet 13.4% at Top Firm: Profits per equity partner (PEP) at a major Big Law firm tanked 13.43% last year, according to new ALM data. The sharp drop raises serious questions about firm strategy and market headwinds.
• EFF to UN: Gaza Journalist Deaths Surge Post-Oct 2023: The digital battleground for truth is heating up, and the frontline is now Gaza. EFF’s stark submission to the UN paints a grim picture of escalating press freedom violations.
• Financial Censorship: A New Frontier?: Forget bots and algorithms for a second. The real frontier of AI and online control might just be your bank account. Former EFF Director Rainey Reitman’s new book, Transaction Denied, pulls back the curtain on financial intermediaries quietly becoming gatekeepers of expression.
• Biglaw’s AI Nightmare: Fake Cases, Real Disasters: AI promised efficiency. It’s delivering fake case law and existential dread instead. Biglaw’s reckoning has begun.
• Claude for Word Flops in Legal AI Test [Ivo Report]: Turns out your fancy general AI chatbot might not be ready for prime time when it comes to the nitty-gritty of legal contracts. A new test shows Claude for Word, well, it flopped.
• Lawyers & AI: Is courtesy dead or just postponed?: The esteemed Legal AI Beat isn’t about AI this time. It’s about lawyers. And a simple courtesy. Why is asking for an adjournment becoming a tactical battlefield?