Ask The PhD Community

HigherEd AI Daily

June 15 – AI Export Controls Reach Campus Research Tools

Monday, June 15, 2026

Today's edition tracks three developments with direct implications for how campuses govern AI tools, design curricula, and think about what it means to transfer knowledge in an age of conversational AI.

TLDR AI + The Rundown AI | POLICY

Anthropic Shuts Down Fable and Mythos Globally Under U.S. Export Control Directive

Anthropic has shut off worldwide access to Mythos 5 and Fable 5, its two newest and most powerful AI models, following an export control directive issued by the Trump administration's Commerce Department. The directive restricts use of the models to U.S.-citizen users only, citing intelligence concerns about a China-linked group that may have accessed the Mythos model. Anthropic stated that a complete shutdown is the only viable path to compliance while the directive is in effect; all other Anthropic models remain available.

The corporate and political context adds considerable complexity. Anthropic investor Amazon reportedly flagged potential Fable vulnerabilities to government officials, helping trigger the directive. CEO Dario Amodei has long advocated for stronger AI regulation; this action represents government intervention in a form the company did not anticipate. Anthropic characterized the underlying security concern as a "minor and non-universal" jailbreak and noted that similar vulnerabilities exist across other frontier systems, including OpenAI's GPT-5.5.

The move is being described by analysts as "far more than a jailbreak story" and as a preview of how export control frameworks designed for physical hardware and semiconductors may be extended to AI software at any time.

Why it matters for campuses

Institutions that have embedded Fable or Mythos into research workflows, writing support systems, or academic tools will need to pivot immediately. More significantly, this episode signals that AI software now sits within the scope of export control policy; a development that academic technology officers, international programs offices, and research compliance teams should begin tracking in earnest. Campuses with large international student or faculty populations face particular exposure as these policies develop and potentially expand.

Read More

The Rundown AI | GOVERNANCE

China's Universities Have Cut More Than 12,000 Degree Programs to Accelerate AI Transition

China has eliminated more than 12,000 university degree programs over the past five years in a nationally coordinated effort to align higher education with an AI-driven economy. The programs cut span arts, languages, and traditional humanities fields; they are being replaced by technology, engineering, and applied AI disciplines. The restructuring is explicitly tied to government analysis of how AI is reshaping the graduate labor market and national economic competitiveness.

The scale of reform reflects a centrally directed national strategy that differs substantially from the market-driven, institution-by-institution curricular adjustments that characterize most Western higher education systems. Chinese planners are treating the workforce transition as a national security and economic priority rather than leaving adaptation to disciplinary inertia and distributed shared governance processes.

Why it matters for campuses

For U.S. colleges and universities, China's aggressive curricular overhaul raises strategic questions about program relevance, workforce alignment, and the timeline for meaningful AI integration into existing degrees. Academic administrators and department chairs watching enrollment trends in humanities and general education should engage these data points carefully; the pressure to restructure is arriving from multiple directions simultaneously, and faculty governance bodies will need to grapple with these questions on their own terms and timelines rather than wait for a top-down directive.

Read More

TLDR AI | RESEARCH

Has AI Already Disrupted the Market for How-To Information?

Entrepreneur and author Tim Ferriss published a detailed analysis arguing that the market for information-based nonfiction is collapsing into the chatbot. Drawing on his own book sales data and broader publishing industry trend lines, Ferriss examines whether AI assistants have substantially eroded demand for how-to books, explainer content, and practical reference guides; the category that has long served as the commercial engine of trade nonfiction publishing.

The pattern he documents is a structural shift: readers who previously purchased books to learn a skill, solve a problem, or understand a domain are now beginning that journey with a conversational AI query. The path from question to book purchase is no longer reliable, and in many categories it appears to have shortened or ended at the first AI response. The implications extend well beyond publishing to any field built on the transfer of packaged information.

Why it matters for campuses

This trend has layered implications for higher education. Faculty who write practitioner-facing books, departments with certificate or continuing-education programs built primarily on information transfer, and academic publishers all face compounding market pressure. More immediately, instructors whose courses rely on assigned nonfiction reading will want to examine whether students are substituting AI responses for the assigned text; a pedagogical question with significant assessment design, academic integrity, and course architecture implications.

Read More

Tool of the Day

OpenRouter Fusion

OpenRouter Fusion is an API that pools responses from multiple frontier AI models into a single synthesized answer. The system sends a user prompt to several models simultaneously, uses a separate model to evaluate each response, and merges the best elements into one output; with early benchmarks showing the combined result approaches the quality of top-tier single models at a lower per-query cost. For educators and academic researchers navigating the current volatility in AI model availability, Fusion offers high-quality assistance without dependence on any single provider.

Try it: Submit a draft research question or literature review outline to Fusion via the OpenRouter API and compare the synthesized output against what a single model returns; the side-by-side result often surfaces gaps in framing or coverage that no individual model identifies on its own.

Visit OpenRouter Fusion

Have a great learning day!

Dr. Ali Green

Sources for This Edition

TLDR AI (tldrnewsletter.com)
The Rundown AI (daily.therundown.ai)
Ars Technica (arstechnica.com)
South China Morning Post (scmp.com)
Tim Ferriss Blog (tim.blog)
Semafor (semafor.com)

askthephd.com
 | 
askthephd.substack.com
 | 
Unsubscribe

HigherEd AI Daily; Curated by Dr. Ali Green