Bans, Subpoenas, and Hallucinations
· The Fluency Briefing
The Fluency Briefing
Your Guide to What's Happening in AI and Why It Matters to You
Sunday, June 14, 2026
If you're wondering who's actually in charge of AI right now, this week made it clear: nobody, and everybody, all at once. The U.S. government yanked Anthropic's newest models off the market on a Friday evening, 42 state attorneys general served OpenAI with a sweeping subpoena, and KPMG quietly pulled an AI report because, well, the AI they used to write it made things up. Behind every shiny announcement, there's a mess somebody has to clean up.
Today in AI:
- Uncle Sam Pulls the Plug on Claude's Newest Trick - The U.S. Department of Commerce slapped export controls on Anthropic's Fable 5 and Mythos 5 models Friday evening, citing national security concerns tied to a jailbreak vulnerability reportedly flagged by Amazon. Anthropic had to shut both models down entirely since it can't verify users' citizenship. The Verge
- Amazon Played Tattletale - According to the Wall Street Journal, Amazon's cybersecurity research found Fable 5 could be prompted into sharing cyberattack-useful information, and CEO Andy Jassy personally brought those findings to the White House. Translation: your biggest investor just got your product banned. Semafor
- 42 Attorneys General Come for OpenAI - A massive coalition of U.S. state attorneys general served OpenAI with a broad subpoena on June 12, targeting the company's ad practices, data handling, treatment of minors and seniors, and even model sycophancy. This landed just five days after OpenAI filed confidentially for a potentially trillion-dollar IPO. Tom's Hardware
- KPMG's AI Report Gets Caught Hallucinating - KPMG pulled its own report on agentic AI after UBS, the UK's NHS, and others said the claims about their AI usage were flat-out wrong. The culprit? The consulting giant apparently used AI to write a report about AI, and nobody checked the facts. TechCrunch
- Meta's $14 Billion AI Bet Still Isn't Paying Off - A year after poaching Alexandr Wang and top Scale AI engineers, Meta shipped its first proprietary model, Muse Spark, but Wall Street remains unimpressed. The stock is down 18% over the past year, the worst in the megacap group. Now Zuckerberg needs to prove AI can be more than an ad-booster. CNBC
- Google Opens Pinpoint to Everyone - Google's Pinpoint, previously restricted to journalists and academics, is now free for all. It lets you upload up to 200,000 files per collection, transcribes audio and video, and makes handwritten notes searchable. If you drown in documents, this one's worth bookmarking. Fast Company
- India Asks: Should We Build Our Own? - Anthropic's model shutdown reignited debate in India over whether the country's massive AI market can keep depending on U.S.-controlled tech. The timing is especially awkward given Anthropic had just announced a partnership with Tata Consultancy Services to expand enterprise AI there. TechCrunch
- OpenAI Adds Lockdown Mode to ChatGPT - OpenAI introduced a new Lockdown Mode that restricts ChatGPT's access to external connections and web search, specifically designed to guard against prompt injection attacks. Think of it as airplane mode for your chatbot. Whytryai
Today's Takeaway:
Here's the thing about the Anthropic saga: it's not really about one jailbreak. Follow the timeline. In April, Anthropic launched Mythos 5, a model so capable at finding software vulnerabilities that they restricted access to a small group of security-vetted companies. Earlier this week, they released Fable 5, the consumer version with guardrails bolted on. Then Amazon, which owns a major stake in Anthropic, discovered those guardrails could be bypassed through a series of prompts. CEO Andy Jassy brought the findings straight to the White House, and by Friday at 5:21 PM, Commerce Secretary Howard Lutnick had signed an export control directive forcing Anthropic to cut off all foreign nationals, including its own employees. As Semafor reported, the move was partly driven by suspicions that a China-linked group had already accessed the model.
Looking at this through a competitive lens, the picture gets uncomfortable. Amazon is both Anthropic's largest investor and, reportedly, the company whose research triggered the government ban. As The Verge noted, the whole chain of events raises questions about what happens when your financial backer has a direct line to regulators. Anthropic insists the jailbreak was minor and that other models can find the same vulnerabilities. But the damage is done: Fable 5 and Mythos 5 are offline, India is reconsidering its dependence on American AI, and every other frontier lab just got a preview of how quickly government intervention can shut you down.
🧠 AI Trivia - Test Your Knowledge
1. What recent action did Anthropic take regarding its newest AI models, as reported in the news? a) Released them open-source to the public b) Suspended access following a U.S. government directive c) Announced a major partnership with Google for exclusive access
2. Who is widely credited with coining the term "artificial intelligence" in 1956? a) Alan Turing b) John McCarthy c) Marvin Minsky
**3. a) $12 million b) $120 million c) $1.2 billion
Answers at the bottom of the newsletter!
The Bottom Line
The Pattern: This week's stories share an uncomfortable thread: the AI industry is building faster than it can govern itself, and outside forces are stepping in to fill the gap. Whether it's 42 attorneys general, a Commerce Department letter at 5 PM on a Friday, or a consulting firm's own hallucinating report, the gap between what AI companies promise and what they deliver is becoming everyone else's problem.
Why It Matters: If you're a business relying on AI tools, this week is a reality check. The model you're building on today could be pulled tomorrow for reasons that have nothing to do with your use case. And the reports informing your AI strategy might literally be making things up. Trust, but verify, isn't just good advice; it's survival.
Your Move: Before your next AI purchase or integration decision, ask one question: what's my fallback if this tool disappears overnight? If you don't have an answer, that's your project for this week.
📝 Trivia Answers: 1) b - Anthropic suspended access to its newest AI models due to a U.S. government directive, raising questions in the global technology industry. | 2) b - John McCarthy, a computer scientist, coined the term "artificial intelligence" at the Dartmouth Conference in 1956, marking a pivotal moment for the field. | 3) a - Training OpenAI's GPT-3 model is estimated to have cost around $12 million, a substantial investment for its development.
What We're Working On
✨ Founding Cohort Special - 60% Off! - Use code MAF20 to join for just $20/month (regularly $50). Get weekly group sessions & workshops, self-paced courses for all levels, access to tools & templates, challenges with peer feedback, and 24/7 support community. → Join Now
✨ Free 30-Minute AI Consultation - Discover how My AI Fluency can help your business unlock the potential of AI. We'll discuss your goals, explore practical AI opportunities for your industry, and outline clear next steps. → Schedule Free Call
✨ How AI-Fluent Are You? - Test your AI fluency with our interactive quiz. See how you stack up and discover what to learn next. → Take the Quiz
💬 Community | 📞 Book a Consultation | 🌐 Website
Fluently yours, The My AI Fluency Team