Ask The PhD Community

HigherEd AI Daily

May 27 – Curriculum, Equity, and AI’s Moral Turn

Wednesday, May 27, 2026

From curriculum design to hiring equity to a papal ethics document, the questions AI is posing to higher education have never been more urgent or more cross-disciplinary.

The Rundown AI — AI Education

Jensen Huang Pushes Back on the Search for “AI-Proof” Subjects

Nvidia CEO Jensen Huang made headlines this week by challenging the growing trend of parents and students searching for AI-proof academic disciplines. Speaking in Singapore, Huang argued that the skills that mattered before AI will still matter in the age of AI; the more productive question, he said, is not which subjects are safe from automation but how AI can elevate a student’s learning, craft, and purpose. He used journalism as his example, noting that the best practitioners do not merely prepare questions but listen, read the room, and respond with judgment in real time.

Huang invoked the Japanese concept of wabi-sabi, the beauty found in imperfection, to suggest that qualities such as taste, originality, and emotional resonance will become more valued across professional domains precisely because AI handles more of the routine. He also pushed back on narratives linking current job cuts directly to AI, calling the framing “lazy.” Critics were quick to note that more than 80,000 tech positions have already been eliminated in 2026, with companies citing automation in public statements.

Why it matters for campuses

Curriculum committees and academic affairs offices are navigating precisely the question Huang addressed: which competencies to center in an AI-augmented world. His framing, stop chasing AI-proof subjects and start asking how AI elevates the craft, offers a useful heuristic for program redesign. Faculty and deans can engage it critically alongside labor market data and institutional mission; it is not a final answer, but it reframes a conversation that too often collapses into anxiety about replacement rather than inquiry about augmentation.

Read More

The Rundown AI — Access & Equity

Stanford Research Documents Clear Racial Disparities in AI Hiring Tools

A major Stanford-led study has produced some of the most direct evidence to date that AI-powered screening tools create measurable racial inequity. Researchers analyzed 4 million job applications across 156 employers and found that widely deployed AI assessment platforms disproportionately screened out Black and Asian applicants; 10.62% of positions showed adverse impact against Black candidates, and 5.32% against Asian candidates. The study focused on Pymetrics, an AI assessment tool used broadly across industries.

A compounding structural problem emerged from the research: because 42 models are shared across employers using the platform, a candidate rejected by one organization may be systematically rejected by others using the same underlying model. The researchers found that 4% of applicants who applied to 10 positions were rejected from all of them at rates higher than would be expected if employers were making independent decisions. The data spans 2018 to 2022, and the researchers caution that their findings may not generalize to today’s LLM-based hiring tools, which differ architecturally; the underlying structural risk, however, is not architecture-specific.

Why it matters for campuses

Institutions with commitments to equitable hiring that also use AI-assisted screening tools for staff and faculty searches face the risks documented in this study. Procurement officers and human resources directors should audit whether vendor platforms use shared models; legal and compliance teams should ask what adverse impact testing has been done and on which populations. The risk is not hypothetical, and “we use a third-party tool” does not insulate an institution from responsibility under Title VII or analogous frameworks.

Read More  (subscription may be required — full paper: algorithmichiring.github.io)

TLDR AI — Governance & Policy

Pope Leo XIV Publishes a Formal Encyclical on the Ethics of Artificial Intelligence

Pope Leo XIV has issued a papal document on the ethics of artificial intelligence, representing one of the most substantive interventions by a major moral institution in the ongoing governance debate. The encyclical addresses the environmental cost of AI infrastructure, the risks posed by algorithmic systems that make consequential decisions about people’s lives, and the way AI concentrates power in the hands of those who already command the most resources. Analyst Simon Willison, whose widely read commentary is linked in the coverage, describes the document as approachable and technically grounded, noting that it requires no Catholic or theological background to engage with meaningfully.

The encyclical arrives at a moment when regulatory bodies, professional associations, and university ethics committees are developing their own AI governance frameworks. It joins UNESCO’s Recommendation on the Ethics of AI, the Montreal Declaration for Responsible AI, and the European AI Act as a significant non-governmental moral framework that institutions across sectors are consulting as they draft internal policy.

Why it matters for campuses

For faculty in philosophy, theology, law, public policy, and computer science, the encyclical is a primary source worth assigning. For administrators developing or revising AI governance documents, it offers a moral framework grounded in concepts of human dignity, equity, and ecological responsibility that translates across traditions. It also signals something broader: AI governance has become a cross-disciplinary and cross-cultural conversation, and campus leaders who engage it only through legal compliance or technical policy miss the full dimension of what institutions are being asked to address.

Read More

TLDR Dev — Research & Wellbeing

AI-Assisted Workflows Are Producing a New Kind of Cognitive Burnout

A detailed analysis from Evil Martians argues that AI-assisted development is producing a qualitatively new form of cognitive burnout among software engineers. The core finding is that reviewing AI-generated output is mentally exhausting in ways that differ from the fatigue of creative problem-solving; it demands sustained vigilance and critical judgment but provides none of the compensating satisfaction of craft. Engineers report processing far more content than before while feeling less connected to the work being produced.

The authors recommend that practitioners set deliberate boundaries on how much automation they accept, preserve time for tasks requiring manual skill and interpersonal communication, and treat human-centric capacities as something to be actively maintained rather than assumed. The piece runs to 20 minutes and draws on practitioner accounts as well as cognitive science literature on attention and overload.

Why it matters for campuses

The pattern documented here maps directly onto what is emerging in academic workflows. Faculty who grade AI-assisted student work, instructional designers building AI-enabled coursework, and graduate students who offload writing or analysis to AI tools may all be navigating the same tradeoff: volume and speed against cognitive engagement and skill development. Institutions developing AI policies focused exclusively on academic integrity are missing an equally important dimension; the wellbeing implications of AI-augmented labor are real, and campuses would do well to build that conversation into faculty development and student advising before burnout becomes visible in attrition data.

Read More

Tool of the Day

Anytype

Anytype is an encrypted, local-first knowledge and collaboration platform designed for individuals and teams who want full control over their data. It uses on-device encryption with user-held keys, meaning the company and third parties have no access to stored content. The platform includes a block-based editor, private relational databases, and built-in chat, making it well-suited for researchers, faculty, and instructional designers who handle sensitive documents, IRB materials, grant records, or collaborative writing that should not pass through commercial cloud servers.

Try it: Create a private research notebook for an ongoing project, using Anytype’s linked database feature to connect literature notes, methodology drafts, IRB documentation, and collaborator meeting notes into a single encrypted workspace that stays on your device.

Visit Anytype

[Dr. Green’s closing note to be added before sending.]

Dr. Ali Green

Sources for This Edition

The Rundown AI (therundown.ai)
TLDR AI (tldrnewsletter.com)
TLDR Dev (tldrnewsletter.com)
Business Insider (businessinsider.com) — Jensen Huang on AI education
Financial Times (ft.com) — Stanford AI hiring bias study
Stanford Algorithmic Hiring Study (algorithmichiring.github.io)
Simon Willison’s Weblog (simonwillison.net) — Pope Leo XIV AI encyclical
Evil Martians (evilmartians.com) — AI cognitive burnout

askthephd.com
 | 
askthephd.substack.com
 | 
Unsubscribe

HigherEd AI Daily; Curated by Dr. Ali Green