Category: Higher Ed

AI’s rapid capability leap promises transformative productivity—from research and corporate automation to search and teaching—prompting claims of AGI. Simultaneously, security, integrity, and governance risks (malicious packages, cheating, privacy, supply‑chain concerns) demand institutional guardrails, human judgment, and regulation.

• The Verge: Nvidia CEO Jensen Huang says ‘I think we’ve achieved AGI’ (Mar. 23, 2026)
  On Lex Fridman’s podcast, Nvidia CEO Jensen Huang said, “I think we’ve achieved AGI,” claiming AI can start, grow, and run major companies. 
• NY Times: Is Taste the One Thing A.I. Can’t Replace? (Mar. 22, 2026)
  Tech workers are trying to cultivate “taste,” a human judgment guiding AI, as a skill that could keep them indispensable. They teach courses, host dinner salons, and debate whether taste is teachable, subjective, or merely a buzzword.
• Out of Sample: A Second Industrial Enlightenment (Mar. 23, 2026)
  AI breakthroughs suggest machines can accelerate analyses, proofs, and experiments, approaching Nobel-level research in some fields. But limits in interpretation, verification, and incentives mean institutions are needed to sustain growth.
• WSJ: AI Is Rewriting the Old Rules of Google Search and SEO (Mar. 22, 2026)
  AI-powered search is changing SEO, so companies must become trusted sources by engaging customers, structuring content, and adding contextual details.
• Simon Willison: Malicious litellm_init.pth in litellm 1.82.8 — credential stealer (Mar. 24, 2026)
  LiteLLM on PyPI hid a base64 credential stealer that ran on install, stealing SSH keys, cloud credentials, wallets, and histories.
• The Chronicle of Higher Education: AI Broke College Assessment. One University Believes It’s Got a Fix. (Mar. 23, 2026)
  U.S. colleges mostly leave AI policy to individual professors, creating a confusing patchwork, straining faculty, and harming integrity. 
• Inside Higher Ed: Canvas Unrolls AI Teaching Agent (Mar. 23, 2026)
  Canvas launched IgniteAI to automate routine tasks, free faculty for mentoring, and include guardrails to prevent automated grading. Critics warn agentic AI may enable cheating, erode student-faculty ties, and push higher education toward larger classes.
• The Chronicle of Higher Education: Is AI Making Us Stupid? Cal Newport Is Worried. (Mar. 12, 2026)
  Cal Newport warns AI threatens to outsource hard thought, and urges universities to cultivate cognitive fitness, protect deep work, and ban AI-written student work. 
• NY Times: In N.Y.C. Classes, Teachers Can Use A.I. to Plan but Not to Assign Grades (Mar. 24, 2026)
  New York City released an A.I. playbook letting teachers use A.I. for lesson planning, research, and drafting, but banning it for assigning grades or disciplinary decisions. It stresses guardrails, data privacy, teacher judgment.
• The Register: US bans any new consumer-grade routers not made in America (Mar. 24, 2026)
  The FCC added all foreign-made consumer routers to its Covered List, banning approval of new models, while allowing previously authorized devices. It cites national security, critics say most routers are made abroad, and DoD, DHS, and conditional approvals are allowed.

• Anthropic: Introducing Sonnet 4.6 (Feb. 17, 2026)
  Claude Sonnet 4.6 delivers major upgrades in coding, computer use, long-context reasoning, agent planning, knowledge work, and design, with a 1M-token context window.
• Sar Haribhakti: GitHub Commits & Claude Code (Feb. 5, 2026)
  “4% of GitHub public commits are being authored by Claude Code right now. At the current trajectory, we believe that Claude Code will be 20%+ of all daily commits by the end of 2026.”
• The Chronicle of Higher Education: David Brooks: ‘We’re Part of the Problem’ (Feb. 17, 2026)
  “Brooks argues the Ivy‑League–driven American meritocracy warped higher education, prioritized résumé virtues over moral formation, hardened social lines, and helped produce political polarization (including Trump).”
• Axios: Anthropic’s Claude Opus 4.6 uncovers 500 zero-day flaws in open-source code (Feb. 5, 2026)
  Anthropic’s Claude Opus 4.6 found over 500 unknown, high-severity open-source vulnerabilities with little prompting.
• Mitchell Hashimoto: My AI Adoption Journey (Feb. 4, 2026)
  A five-step journey adopting AI agents: drop chatbots, reproduce work with agents, run end-of-day agents, outsource tasks, engineer safeguards. It boosted focus, efficiency, and control.
• Andrew Yang: The End of the Office (Feb. 16, 2026)
  Some dire prognostications here: AI is rapidly automating white-collar work, threatening massive layoffs, bankruptcies, housing declines, and social unrest. 
• NY Times: Why an A.I. Video of Tom Cruise Battling Brad Pitt Spooked Hollywood (Feb. 16, 2026)
  A 15-second AI-generated clip of Tom Cruise fighting Brad Pitt, made with ByteDance’s Seedance 2.0, looked cinema-quality and spread widely. Hollywood responded with outrage, cease-and-desist demands, and renewed fears.
• Inside Higher Ed: AI Can Raise the Floor for Higher Ed Policymaking (opinion) (Feb. 5, 2026)
  AI helped committees gather and compare policies, research, and feedback, shifting discussions from anecdotes to evidence. It broadened options, improved implementation, and supported ongoing, data-driven policy review.
• WSJ: How to Stay Sane in the AI Skills Race (Feb. 4, 2026)
  Don’t panic, assess your role, choose targeted training, and build a portfolio. Certificates alone won’t impress employers.
• NY Times Opinion: What if Labor Becomes Unnecessary? (Feb. 4, 2026)
  Economists debate A.I.’s labor impact, noting current employment data are inconclusive, while massive A.I. investment could lock in major disruption. 
• Tyler Cowen: “You see tech and AI everywhere but in the productivity statistics” (Feb. 16, 2026)
  Brynjolfsson notes a 403,000 payroll downward revision, while Q4 real GDP rose 3.7%. That mix—more output, less labour—signals productivity growth. He projects 2.7% US productivity for 2025, nearly double the decade’s 1.4%.
• Austin Vernon: Speed Can Reindustrialize America – Austin Vernon’s Blog (Feb. 12, 2026)
  US manufacturing makes high-volume, static goods, but struggles with low-volume, custom parts. Long lead times and high white-collar soft costs make small runs unprofitable. Digitized, AI-enabled firms cut soft costs, speed production, and boost resilience, and policy should ease approvals.

Financial

Last month, OpenAI’s shift to a for-profit public-benefit corporation lifted Microsoft’s valuation past $4T. CEO Sam Altman continues to travel the world for funding deals, and October was a busy month for him. Technology contrarian Ed Zitron calculates OpenAI’s cash needs over the next 12 months to be $400B, but Fed Chairman Powell dispels the connection between AI funding boom and Dotcom crash: “I won’t go into particular names, but they actually have earnings.” (Fortune).

Criticism abounded as OpenAI’s CFO opened a can of worms by suggesting government guarantees for data centers, only for Altman to walk these claims back. Critics of OpenAI are raising alarms.

Why does this matter?

If you have a retirement account, you’ll likely care about a potential stock market correction or crash. Aside from that, the financing of AI data centers has tentacles into other companies and industries (Oracle, Google, Nvidia, Meta, Microsoft, power companies, etc.), so downturns and bankruptcies would likely lead to market disruption. For higher education, there are considerations about AI model pricing, and stock market fluctuations can affect giving to non-profit organizations.

…

AI & Productivity

Amazon & UPS announced layoffs at the end of last month. From Amazon SVP, Beth Galetti: “This generation of AI is the most transformative technology we’ve seen since the Internet, and it’s enabling companies to innovate much faster than ever before (in existing market segments and altogether new ones).” Analyst Gil Luria suggests “companies appear to be making the cuts partly to hold their overall profit margins steady while they spend tens of billions of dollars on A.I. infrastructure like data centers. Cutting back on employees is a way to convince shareholders.”

But Luria also notes: “[w]e do think that at some point A.I. tools will allow us to enhance productivity to a point that we’re going to need less labor, but we’re not there yet, not in any significant way.” But another way of thinking about AI & productivity is not merely task augmentation but as something that enables creativity. From developer Aaron Boodman:

“Claude doesn’t make me much faster on the work that I am an expert on. Maybe 15-20% depending on the day. It’s the work that I don’t know how to do and would have to research.

Or the grunge work I don’t even want to do. On this it is hard to even put a number on.

Many of the projects I do with Claude day to day I just wouldn’t have done at all pre-Claude. Infinity% improvement in productivity on those.“

(Emphasis mine)

Why does this matter?

As I mentioned in an earlier post, the potential of a J curve for AI productivity gains is one that some economists suggest. Although productivity gains aren’t yet visible, there is growing anecdotal data to suggest structural changes in work, particularly in visual and technical fields. 

…

AI & Higher Education

Wharton Human-AI Research reported that many enterprises have incorporated AI tools into employees’ daily work and are no longer exploratory in nature.

Higher ed, meanwhile, is not using AI to the same degree. Only 2% of Student Success Leaders say their institutions are very effective at using AI. Their measure is subjective, but the picture is suggestive that AI adoption in higher education is slower than in industry (for good or for ill). Higher ed Leaders are exploring governance and policy, a task likely to be difficult for wrangling fast-moving AI technological advancements. 

What does this matter?

Universities continue to explore using AI, but at a pace slower than industry. There are opportunities for universities to participate in both the conversations framing the use of AI and the practical use of the tools.