Ask The PhD Community

HigherEd AI Daily

May 19 – A Nation Treats AI Literacy as Public Infrastructure

TUESDAY, MAY 19, 2026

Today's edition centers on a question higher education has been circling for two years; what happens when AI access and AI literacy become treated as basic civic infrastructure, and how prepared are our campuses to meet that moment.

The Rundown AI – Policy & Access

Malta Becomes the First Country to Offer Free ChatGPT Plus to Every Citizen Through a National AI Literacy Course

OpenAI and the government of Malta announced a country-wide partnership that will provide free ChatGPT Plus access to any citizen who completes a state-administered AI literacy course. It is the first national-scale deal of its kind, framing AI access as a public utility rather than a consumer subscription, and tying that access to verified completion of a structured curriculum.

The pairing of access with a literacy gate is the meaningful design choice; Malta is not simply distributing tools, it is sequencing access behind formal learning. Other small nations have signaled interest in replicating the model, and the European policy community is watching closely as a possible template for AI literacy benchmarks across member states.

Why it matters for campuses

Universities now have a live national example of an AI literacy credential coupled to tool access. Provosts, CIOs, and centers for teaching and learning should be asking who designs and validates this kind of curriculum on their own campuses, what counts as proof of AI literacy for students and faculty, and whether institutional licensing should follow a similar earn-it model rather than blanket distribution.

Read More

TLDR – Research

A Five Minute Field Briefing on the Last Six Months in Large Language Models

Researcher Simon Willison published a concise survey of how large language models have shifted between late 2025 and mid 2026. He identifies three inflection points; coding agents that are now reliable enough to use as daily tools, a step change in reasoning quality across leading labs, and the maturation of open weight models that run on consumer hardware.

The piece is short by design and is the kind of grounded synthesis faculty and instructional designers can read in one sitting. It avoids hype, names specific releases, and gives a fair baseline for anyone who needs to know what has actually changed before walking into a syllabus revision or a department conversation.

Why it matters for campuses

Most faculty are still calibrating against models that no longer represent the state of practice. Sharing a primer of this caliber with department chairs and curriculum committees gives them a shared factual basis for revising AI policies, assessment design, and academic integrity language without having to follow daily news cycles.

Read More

TLDR – Governance

When Four AI Agents Were Given Their Own Radio Stations, They Developed Distinct Personalities

Andon Labs gave four leading AI models autonomous control of separate online radio stations and tracked how each one ran the operation over an extended period. The models did not converge on a neutral broadcaster style; they diverged into recognizable personalities ranging from buttoned-up news anchor to activist programmer to drifting eccentric.

The experiment is a useful artifact for classroom and policy conversations because it sidesteps abstract debates about AI alignment and shows, in concrete operational terms, how identical instructions can yield very different long-running behavior depending on the underlying model. The report includes transcripts and host commentary that make the divergences easy to discuss.

Why it matters for campuses

For ethics courses, journalism programs, communication studies, and any unit thinking about agentic AI in administrative workflows, this is a teachable case study; long-horizon autonomy is where model differences actually surface. Pair it with your existing readings on bias and content moderation for an immediate seminar discussion.

Read More

The Rundown AI – Tools

A Working Recipe for Generating and Editing 3D Models in Blender Using Plain English Prompts to Claude

The Rundown published a step-by-step walkthrough that connects the open source 3D suite Blender to Claude through the Model Context Protocol, allowing a user to generate, light, and render scenes by typing requests rather than clicking through menus. The setup is documented end to end and runs on a standard laptop.

For instructional designers building course materials, faculty in STEM and the visual arts, and librarians supporting maker spaces, this lowers the barrier to producing custom 3D assets considerably. The same workflow can be adapted to assignments that ask students to describe what they want to see and then reflect on how their prompts shaped the output.

Why it matters for campuses

3D modeling has long sat behind a steep software-skills wall that excluded many disciplines from using it. Workflows like this one open a path for art history, archaeology, biology, and design programs to bring custom 3D into the classroom without first running a Blender bootcamp, and they offer a concrete example of how MCP plumbing is starting to reshape teaching tools.

Read More (subscription may be required)

Tool of the Day

Files.md

Files.md is a free, open source environment for keeping notes, research journals, and task lists in plain Markdown files that stay on your own device. It runs entirely in the browser with no account and no cloud sync by default, and it adds a built in chat interface plus an optional Telegram bot for quick capture from a phone. The local-first design makes it a sensible fit for faculty and graduate students who want a working AI-adjacent notebook without sending student data, draft manuscripts, or grant work to a vendor cloud.

Try it: spend ten minutes today migrating one course's running planning notes into a Files.md vault, tag entries by week and learning outcome, and use the built in chat to ask the tool to summarize where your students struggled most in the past three weeks.

Visit Files.md

[Awaiting closing from Dr. Ali Green. Please replace this bracketed line with your two sentence closing in your own voice before sending.]

Dr. Ali Green

Sources for This Edition

The Rundown AI (therundown.ai)
TLDR (tldrnewsletter.com)
OpenAI (openai.com)
Simon Willison (simonwillison.net)
Andon Labs (andonlabs.com)
Files.md on GitHub (github.com/zakirullin/files.md)

askthephd.com
 | 
askthephd.substack.com
 | 
Unsubscribe

HigherEd AI Daily; Curated by Dr. Ali Green