Four things to know
before your first meeting
The table of contents for a busy week — read this, then decide whether to keep going.
Rep. Suzanne Bonamici introduces the Artificial Intelligence Education and Workforce Readiness Act — Congress's first substantive higher-ed AI bill after years of silence.
Source →A provost's own opinion writing gets flagged as AI-written by the same detector campuses use on students — and Harvard's faculty survey numbers land too.
Source →Microsoft publishes 5 principles for AI in education, with campus case studies and a wage-premium figure for AI skills that beats an advanced degree.
Source →OpenAI's certification push and a16z's tuition-free "SF Academy" both aim at the same target: the path from classroom to credential to job.
Source →Legislation, a scandal,
and a wage number
Four developments with enough substance to change what you do this week.
Congress introduces its first major higher-ed AI bill
Rep. Suzanne Bonamici (D-OR) is introducing the Artificial Intelligence Education and Workforce Readiness Act this week, aiming to mandate AI risk assessments and fund training and credentialing at federally funded colleges — the first substantive federal legislative move on AI in higher ed after years of congressional silence.
- The bill would mandate AI risk assessment for federally funded institutions.
- It creates new grants for faculty training, credentialing, and research on AI's impact on education and equitable access.
- It comes as a 2026 Survey of College and University Presidents found 52% believe the sector "is not responding appropriately to nor prepared to handle the rise of AI."
- Bonamici's office says the bill emphasizes "student-centered solutions for ethical and responsible use of AI in higher education."
If the bill advances, expect federal funding to eventually carry AI risk-assessment reporting requirements — worth flagging to your grants office now rather than after a rule is final.
AI detection got personal: a provost caught by his own tool, and Harvard's numbers are in
Update on the AI-detection coverage from our Sept 15 issue: this week added a named scandal and hard survey numbers, not just more detector debate.
The Dartmouth found that Provost Santiago Schnell's own published opinion writing was flagged as up to 96–100% AI-written by the detector Pangram — including a Washington Post column arguing "a degree should distinguish what students can do independently from what they can accomplish with AI." Separately, a Harvard Crimson faculty survey put real numbers on the distrust for the first time: 88% of faculty say they've seen student work they believe is AI-generated, but only 64% feel confident distinguishing it, and just 12% have reported unauthorized use to the honor council. About a quarter of Harvard faculty fully ban AI, 64% allow it under some policy, and only 3.5% have no policy at all. Separately, University of Wisconsin biology professor Timothy Paustian caught 60 AI-written essays in one 350-student class using hidden prompts last year — but says students have since learned to strip the prompts before submitting.
What this means for your campus: If anyone in your administration writes publicly on AI, assume someone will eventually run it through a detector. Model the disclosure practice you want from students.
The Dartmouth → The Atlantic → Harvard survey →Microsoft's 5 principles, and a 40% wage premium for AI skills
Update on the Microsoft/AFT student-data pact from our Sept 15 issue: this week Microsoft published the full framework the pact sits inside, plus new case-study numbers.
Microsoft released a five-principle framework for AI in education — safety and privacy by design, educator control, learning-centered AI design, institutional system strengthening, and AI skills preparation — plus case studies showing measurable campus adoption. Auburn University's "Teaching with Artificial Intelligence" course now reaches nearly 10,000 faculty across more than 90 institutions. Babson College built an interdisciplinary AI lab ("The Generator") integrating Azure OpenAI tools and Copilot across every faculty member and student. University of South Florida's student-ambassador program reports 55–60% time savings on some administrative tasks. Underpinning all of it: Microsoft's AI Economy Institute finds AI-related skills now carry wage premiums up to 40%, more than an advanced degree.
What this means for your campus: The wage-premium figure is a concrete argument for your curriculum committee — AI literacy requirements now have a labor-market number behind them, not just a pedagogical one.
Microsoft → Complete AI Training →OpenAI is building a pipeline from classroom to job, and universities are noticing
A Guardian op-ed this week sharpened a growing worry in higher ed: OpenAI's new certification programs explicitly aim to "connect learning, certification and real economic opportunity in one clear journey," laying groundwork for an upcoming OpenAI Jobs Platform that would let the company compete directly with the credentialing role universities have long held. OpenAI is hiring students as campus ambassadors to extend its reach inside universities. Separately this week, Andreessen Horowitz launched a tuition-free "SF Academy" fellowship for high-school grads and early college students, backed by Anthropic, Google, Meta, and other AI companies — no traditional grades, tests, or homework, with a two-year alternative degree program planned pending regulatory approval for fall 2028. Critics quoted in coverage (echoing scholar Ella Hafermalz) warn OpenAI "could soon be selling young people a one-stop-shop for learning, credentials and a job, all heavily dependent on its tools."
What this means for your campus: Career services and provost offices should treat this as a competitive signal, not a partnership opportunity by default — be ready to articulate what a credential from your institution offers that a company-run pipeline doesn't.
The Guardian → Hoodline →What the studies say,
what colleagues tried
Two findings and two classroom moves worth borrowing.
From the Research
We barely study AI's effects on K-12 learners
What was studied: A review of 173 international peer-reviewed studies on generative AI in education (Storrøsæter et al., with colleagues from Czech universities).
What they found: 80% of study participants across the reviewed literature were university or college students — meaning almost nothing is rigorously known about generative AI's cognitive and developmental effects on K-12 children and teens, even as school-level adoption accelerates.
What you could do with this: If you supervise pre-service teachers or run K-12 partnerships, name this evidence gap explicitly rather than assuming your own classroom findings generalize downward to younger students.
Source →AI is already shaping how prospects pick a program
What was studied: UPCEA and Search Influence surveyed 1,057 prospective continuing- and online-education students (fielded May 5–14, 2026) on how they research programs.
What they found: 54% of prospects now use AI tools during program research, though search engines remain the dominant starting point (80%); prospects narrow to three or fewer programs before applying, and AI-generated search overviews increasingly shape that shortlist.
What you could do with this: Check what an AI search overview currently says about your program — if it's wrong, thin, or missing, you may be getting quietly cut before a prospective student ever visits your site.
Source →How Others Are Doing It
She surveys her own students before writing policy
Who: Annette Vee, PhD, Associate Professor of English and Director of Composition, University of Pittsburgh.
What she did: Surveyed her own students directly about their AI use and attitudes before redesigning course policy, tracking how answers changed semester to semester since ChatGPT launched.
What happened: Student use and comfort with AI shifted fast and unevenly — early surveys found near-zero use, later ones found roughly half of students comfortable with AI "fixing errors, formalizing tone" but split on generative use for ideas.
What you could borrow: Survey your own students before writing (or rewriting) an AI policy — assumptions about what students actually do with AI are frequently out of date within a single semester.
Source →A tiered governance structure across campuses
What they did: Adopted Administrative Policy Statement 6012, "Responsible Use of Artificial Intelligence," effective this August, building on four existing campus-level AI groups.
What happened: Faculty Council added a dedicated special meeting on AI; President Todd Saliman is now assembling a systemwide AI group from campus chancellor and governance nominations.
What you could borrow: A tiered structure — campus-level groups feeding into one systemwide body — lets a multi-campus system coordinate policy without flattening real differences between campuses.
Source →- If a provost's own writing gets flagged AI-written by the same tool used on students, how much weight should your integrity process actually put on a detector score?
- OpenAI wants to own the pipeline from classroom to credential to job. If a student could get that same journey from a company instead of your institution, what would your degree still offer that theirs doesn't?
Six fast items,
two dates, three actions
Scannable in under a minute — plus what's coming and what to try this week.
-
1MIT's August AI report keeps rippling Education consultants now expect peer institutions to publish their own "answering reports" by spring 2027.
-
2Google's AI Educator Series adds real college credit Undergraduate credit via College Unbound, graduate credit via ISTE and Dominican University, tied to a Sept 19 "badge-a-thon."
-
3Penn State World Campus launches new AI credentials Funds nine faculty-led AI-literacy projects across its online programs, including a systems-engineering master's and a studio-art course.
-
4States push schools to use AI while leaving teachers to police cheating alone A Stacker/K-12 investigation finds patchwork district policies and no real support for detection.
-
5Ethan Mollick's "The Overhang" A widely discussed essay argues four human advantages — deep knowledge, wide knowledge, taste, and agency — stay scarce even as AI capability grows.
-
6Seoul AI-writing experiment: higher scores, weaker recall A JoongAng Ilbo classroom experiment found AI-assisted essays scored higher but showed weaker recall of students' own arguments.