September 2026 - Research Updates and News, AI Adoption, Learner Protections, and More

September 2026, Eye on Research

Commentary

Moving Faster than the Evidence

Fall is often the time we discover whether the plans we made over the summer hold up. This year, many of those plans involve artificial intelligence (AI). Read side by side, several studies in this month’s roundup show a pattern worth reflecting upon. Higher education is adopting AI faster than we can learn whether it works, or who bears the cost when it doesn’t.

Consider three recent studies.

Researchers interviewed by Inside Higher Ed describe independent evidence of AI’s learning benefits as close to nonexistent. The tools change every few months, while rigorous studies take years.
A new report from Student Defense’s SHAPE-AI initiative finds protections for learners have not kept pace with AI use in admissions, advising and teaching.
A working paper from Census-affiliated economists finds graduates in the most AI-exposed majors entered a job market that resembles a recession, with fewer job offers and lower starting pay than their pre-ChatGPT peers.

None of this surprises me because AACRAO’s own research twice showed the same gap last year. In our LEARN Commission Green Paper on AI and credit mobility, nearly everyone saw AI’s potential to improve credit evaluation, yet few were using it. Our AI in Academic Operations survey told a similar story, with broad optimism and limited adoption. I read those gaps as caution, and much of that caution is healthy. This month’s research is a reminder that caution by itself is not a plan.

For registrars and admissions professionals, AI questions are operational ones.

Does the AI tool apply the articulation rule correctly?
Who reviews the output before a learner sees it?
What happens when it’s wrong?

The SHAPE-AI report singles out advising as particularly risky because incorrect guidance can cost a learner time and money. That same risk applies to a transfer-credit evaluation or a degree audit.

When evidence is thin and the sales pitch is confident, it’s tempting to let the most certain voice in the room make the decision. Consider these questions as your institution weighs its next steps.

How will you know whether an AI tool is working on your campus?
Have you defined the criteria before launching?
When an AI-assisted decision reaches a learner and it’s wrong, who owns the error?
Are we preparing learners in AI-exposed fields for a labor market that has already shifted beneath them?

We don’t need to wait for perfect research to act responsibly. We can build our own evidence. This means defining success up front, keeping people in the loop on decisions with real consequences for learners and sharing what we learn with one another.

If your institution is piloting AI in academic operations or credit evaluation, and tracking the results, I would like to hear from you. Please contact me at wendyk@aacrao.org.

AACRAO Research Updates

Chief Enrollment Management Officer Career Profile

The Chief Enrollment Management Officer (CEMO) Career Profile survey will deploy in October 2026. This will be the fourth iteration of this career profile series for the CEMO position. If you hold the CEMO position, please watch your inbox for this survey. We welcome and appreciate your participation.

We have deployed this survey every 3 years since 2017. Moving forward, we will repeat the survey every 5 years.

3-Year Bachelor’s Degrees 60-Second Survey

Three-year bachelor’s degrees have moved from a talking point to a real option on a growing number of campuses. We want to know what that looks like from where you sit. In November 2026, we will deploy a brief (2-to-3-minute) survey on the use of, and interest in, 3-year bachelor’s-degree programs. The survey covers both reduced-credit degrees, most commonly 90 to 96 credit hours, and the standard 120 credits delivered on a compressed calendar.

Whether your institution already enrolls learners in one of these programs, is building one, is still talking it over or has decided against it, we want to hear from you. The questions follow your answers, so you will only see the ones that apply to you. We are seeking many answers.

How do these degrees show up on the academic record and transcript?
When was the registrar’s office brought into the planning?
Which operational pieces have been hardest to get right?
How does your institution treat a 3-year degree if that graduate applies for graduate or professional program admission?

Please watch your inbox in November. The survey takes about 3 minutes. One response per institution is preferred.

Three AACRAO Books Coming in 2027

We have been busy on the publications side. Three AACRAO books are being revised based on AACRAO research and AACRAO competencies for 2027. All three publications will be ready in time for the 2027 AACRAO Annual Meeting to be held April 4-7, 2027 in Long Beach.

Learner Records Management: Retention, Disposition and Archive (Ninth Edition) updates the 2019 records-management guide. It covers what to keep, for how long and what to do with records when their time is up.
The Academic Record and Transcript Guide replaces the 2020 edition. It includes updated recommendations on what belongs on the academic record and the transcript, along with new guidance on digital credentials and noncredit learning.
The Registrar’s Guide: Foundations and Evolving Practice is organized around the AACRAO Competency Set 2026. It is written for new and experienced registrars who want to see where the profession is headed.

These books draw on years of AACRAO research and member experiences. Many professionals reviewed drafts and shared how things really work on their campuses. We’re grateful for that help, and we look forward to putting the finished books in your hands at the Annual Meeting.

Current Higher-Education Research and Related Topics

What New Research Reveals About Learner Success
Inside Higher Ed, Aug 18, 2026

This roundup pulls together four separate studies that challenge common assumptions about what keeps learners enrolled. The patterns across all four studies reveal colleges often have support in place, but learners don’t know it exists or can’t access it quickly.

A Trellis Strategies analysis found financially vulnerable learners report much higher rates of anxiety, depression, and loneliness. Nearly 25% do not know whether their school offers counseling.
The Hope Center at Temple University found the share of learners receiving emergency aid dropped from 44% in fall 2021 to 4% in fall 2025. This occurred during a time when more emergency-aid programs launched during that span.
Michigan State researchers now estimate over 280,000 autistic learners are enrolled in U.S. colleges, three to five times higher than earlier counts. This suggests schools have been undercounting this group for years.
A DeBruce Foundation study found learners with an understanding of how their skills map to career paths, called career literacy, consider roughly three times more job options and earn about $40,000 more annually than peers who lack it.

An Earn-While-You-Learn Path into Teaching
Inside Higher Ed, covering a New America report, Aug 21, 2026

Teacher-degree apprenticeships let aspiring educators earn a paycheck while completing their degree and certification. New America believes the model could help ease the K-12 teacher shortage. These programs have quickly expanded and now run in 157 sites across 44 states and the District of Columbia. The apprentices are different from typical undergraduate learners and the current teaching workforce.

About 6,275 K-12 teacher apprentices are active nationwide; 75% are between ages 25 and 54.
Apprentices are more diverse than current teachers–18% identify as Black and 19% as Latino, compared to 6% and 9% of the existing K-12 teacher workforce.
Much of the early growth was driven by pandemic-era federal funding that is now expiring; programs are shifting toward Pell Grants and prior-learning credit to hold down costs.
Researchers flagged mentorship as underdeveloped territory because coaching an apprentice takes different skills than supervising a traditional student teacher.

New Paper Takes Aim at “Meritocratic” Admissions Policies
Inside Higher Ed, August 27, 2026

A new National Bureau of Economic Research (NBER) working paper by Zachary Bleemer and Jesse Rothstein delves into decades of federal data to show who actually attends selective colleges. It pushes back against claims that top schools are quietly lowering the bar for diversity. The paper finds well-resourced learners, mostly from higher-income families, still dominate enrollment at the wealthiest institutions.

About 65% of learners at Ivy League and top-50 private schools come from the top income quintile.
Over 50% of learners at state flagships come from the top income quintile.
Per-learner instructional spending rose 29% for every 100 points of average SAT scores among enrolled learners, as of 2019.
Lower-scoring learners, who tend to be limited-income, land at schools that spend less per learner.
Less-selective public and 2-year colleges show income distributions closer to the national average.

How Should AI Shape Teacher Education?
Education Week, August 26, 2026

The American Association of Colleges of Teacher Education (AACTE) released a new framework for how teacher-prep programs should handle AI. These programs train over 90% of new K-12 teachers. The framework tries to balance practical AI skills with room for teachers who have ethical objections to the technology.

AACTE represents more than 500 college and university teacher-ed programs.
The framework covers technical skills, pedagogy and ethical judgment for using and teaching AI.
It allows some aspiring teachers to opt out of AI use on ethical grounds and still be certified.
States may find the opt-out provision tricky to implement in practice.

Higher Education Price Index Forecast, August 2026
Commonfund Institute, August 31, 2026

Commonfund’s latest higher education price index (HEPI) forecast pegs inflation for colleges and universities in the United States at 3.5% for fiscal year 2026, based on data through late August. This index tracks cost growth specific to higher education operations, which tends to run differently from general consumer inflation.

The “Nonexistent” Research on AI’s Benefits for Education
Inside Higher Ed · September 14, 2026

Colleges are pouring money into AI tools, but independent research to back up claims about learning benefits barely exists. Experts state that rigorous studies take years to design and run, while AI tools change every few months, making most findings outdated before they’re published.

MIT’s Justin Reich calls the evidence base “almost nonexistent” and says federal and philanthropic funding for large randomized trials has dried up.
Researchers believe studies should focus on AI features across tools, rather than individual products, because 1,000 AI tools are already on the market.
A preprint from an Ivy Tech Community College researcher found several existing peer-reviewed studies claiming AI benefits rely on flawed statistical methods.
The limited research that does exist suggests AI helps learning only with intentional design; used without guidance, it can encourage harmful cognitive offloading.

Report: Learner Protections Haven’t Kept Up with Higher Education’s Adoption of AI
Inside Higher Ed · September 18, 2026

A new report from Student Defense’s SHAPE-AI initiative reveals colleges have rushed to adopt AI across admissions, advising, and teaching without building corresponding safeguards. The authors urge institutions to go slowly and ask hard questions before rolling out new AI products.

Key risk areas include bias in admissions and financial-aid targeting, data privacy, loss of learning community, and lack of transparency about AI use.
AI in advising and counseling carries particular risk of exposing learners to incorrect guidance with real academic or financial consequences.
The report indicates AI errors, such as hallucinations, can occur even when a system appears to be working normally.
U.S. Congresswoman Suzanne Bonamici, responding to the report, called for clearer guardrails so AI expands opportunity, rather than widening gaps.

Report: AI Created Recession-Like Job Market for Some New Graduates
Inside Higher Ed · September 22, 2026

A new working paper from economists affiliated with the U.S. Census Bureau finds graduates in AI-exposed majors are facing labor-market conditions as tough as a major recession. Researchers compared outcomes before and after ChatGPT’s late-2022 launch across 10 major categories ranked by AI exposure.

Computer science, accounting, journalism and engineering graduates rank among the most AI-exposed.
Nursing and education are among the least AI-exposed programs.
Compared to the pre-ChatGPT era, initial employment likelihood for the most exposed majors fell 5%. Beginning earnings dropped 13%.
About 50% of the earnings decline comes from lower pay within the same industries; the other 50% comes from graduates shifting into lower-wage sectors, like retail and food service.
The effects fade somewhat as graduates gain time in the workforce, but researchers say it’s still unclear whether these workers will fully recover long-term.

The Ivy League Brought Back Test Scores, but 90% of Colleges Still Do Not Require Them
Inside Higher Ed · August 28, 2026

New data from FairTest shows the return-to-testing narrative is mostly a story about a handful of elite schools, not a broad trend. Across 2,247 admissions policies for 2026-27, only 7% of colleges require test scores such as the SAT or ACT, while 86% stay test-optional and 4% are test-free. The policy at the remaining colleges could not be confirmed.

Eleven colleges added new testing requirements in the 2026-27 application cycle, while 10 others dropped theirs. The total number requiring scores rose by only one school nationwide.
Among the top-100-ranked institutions, 11% require scores versus 82% who state tests are optional. The remaining 7% were test-free.
Columbia and Princeton remain optional in 2026-27 but will require scores in 2027-28, while Duke and the University of Chicago plan to stay optional indefinitely.
Public flagships mostly hold the line on testing. Universities in Florida, Tennessee and Ohio, and Purdue and the University of Texas, Austin require scores systemwide or campuswide. The University System of Georgia has expanded its requirement from 3 to 7 campuses.
Even without a requirement, more applicants are submitting scores–51% did in the 2025-26 cycle, the first majority since 2020. Counselors state grade inflation makes admissions offices want that extra data point.

Data Shows Tennessee Promise Boosts College Enrollment and Attainment
Inside Higher Ed · August 28, 2026

A new NBER working paper finds Tennessee’s free community college program, the first statewide program of its kind nationally, delivered real gains in enrollment and degree completion for a modest price tag. Researchers drew on 2000-2023 federal postsecondary data and 2000-2024 Census survey data to track outcomes since the program’s 2015 statewide launch. The program grew from a single-county pilot in 2009 to full statewide eligibility by 2015.

College enrollment among 18-to-19-year-old high school graduates rose 5.4 percentage points, or almost 10 percent of their baseline rate, with somewhat larger effects for nonwhite students and those living outside metropolitan areas.
Two-year institutions saw a 15 percent increase in associate’s degrees conferred and a 24 percent increase in transfer-out learners. Associate’s degree attainment rose 2.9 percentage points among 21-year-olds. However, two-year graduation rates fell by roughly 21 percent of the comparison mean, consistent with more academically marginal learners enrolling.
The paper largely eases the worry that free community college pulls learners away from 4-year degrees. It does find evidence of initial diversion into two-year pathways, including an 8 percent increase in transfer students at four-year institutions.
Promise-eligible individuals overall were about 2.7 percentage points more likely to hold a bachelor’s by age 24, driven by women, who were 4.1 percentage points more likely to hold one by ages 23–24. The authors describe the bachelor’s effect as imprecise.
Promise raised wage and salary income by about 7 percent over ages 21–24, though the authors urge caution on that figure. They found no increase in labor force participation, but did find delayed family formation: eligible individuals were about 27 percent less likely to marry before 21 and 17 percent less likely to have a child in the household by 22.
The program, which covers any gap between tuition and fees and a student’s existing grant aid, averaged less than $1,000 per participant per year. The authors estimate it pays for itself under reasonable assumptions about returns to college.