Gen AI Transparency on Federal Regulations
Finally a way to track hundreds of public comments in federal rulemaking process, focusing on institutional accountability

Was this forwarded to you by a friend? Sign up, and get your own copy of the news that matters sent to your inbox every week. Sign up for the On EdTech newsletter. Interested in additional analysis? Try with our 30-day free trial and Upgrade to the On EdTech+ newsletter.
It turns out that gen AI changes the game in analyzing complex and disorganized federal data to add real transparency into the rulemaking process. In near realtime.
The One Big Beautiful Bill (OBBB) provisions on loan limits and accountability will have a huge impact on higher education programs. The rulemaking is in progress, and now you can understand the discussion in realtime and in human terms.
2023: TPS and OPM Guidance
In 2023 when I wanted to review the public comments for the Third Party Servicers (TPS) expansion guidance and the potential rescission of the OPM rev share guidance, it took a lot of manual work. You could do a bulk download of public comments, but a lot of those comments were actually made in attached files. We (by we, I mean my daughter) manually referred to the PDF attachments and classified responses to get a sense of how many and which commenters supported or were against TPS expansion and OPM rev-share elimination. And after that analysis (which was the only public analysis of comments available), we got a rough sense.

2025: Student Loans and Institutional Accountability
Today, the game has changed. Let’s take a look at the ongoing public comment period on the US Department of Education’s (ED’s) planned negotiated rulemaking to implement the elements of the One Big Beautiful Bill into specific rules for colleges and universities. And let’s focus on the student loan changes and institutional accountability provisions.
To cut to the chase, here is an AI-produced podcast from Google’s NotebookLM discussing the 557 public comments available to date.
The podcast discusses how the "One Big Beautiful Bill Act" (HR1) has significantly reformed federal student aid, particularly by eliminating the Grad PLUS program and altering loan limits for graduate students, which takes effect in 2026. The core of the conversation revolves around the Department of Education's crucial task of drafting regulations for this new law, especially the "do no harm" provisions related to program-level earnings. The episode highlights numerous public comments from professional organizations and individuals, making urgent cases for fields like occupational therapy, public health, and physician assistant programs to be classified as "professional degrees" to avoid drastically reduced borrowing caps. These comments emphasize the potential for severe workforce shortages and limited public access to essential services if these vital professions are not adequately supported. Additionally, concerns are raised about the broader impact on equity for vulnerable student groups, the rising cost of tuition, and the need for greater accountability from educational institutions to ensure students receive a valuable education that enables loan repayment.
The discussion even pulls in specific comments to illustrate concepts [starting at 4:01].
They're all basically saying, "Look, our program programs meet the spirit the intent of a professional degree and excluding us they would have severe consequences for our workforce for the public it directly impacts students’ ability to finish these programs.
Let's start with occupational therapy OT. We saw a ton of submissions there Mary Harden, Baylor, Drexel, MGH Institute, TCU, and UW Madison, many others, they seem very aligned.
Oh completely aligned across the board they're hitting three core points about OT graduate programs.
Okay what are they?
One They lead directly to professional lensure and national certification and that's required in all 50 states to practice.
Right?
Two, they demand really substantial graduate level training, usually two, sometimes three years beyond a bachelor's.
Okay?
And three, they are the prerequisite for practicing in a federally recognized health profession. They emphasize how vital OT's are helping people with daily activities, recovering from illness, injury, improving quality of life across the lifespan.
NotebookLM also organizes the myriad comments into a useful mind map.

ChatGPT, for its part, answered my question about how the public comments can be interpreted given that OBBB is written into law, and ED must create regulations within those boundaries. Here is a subset of the response.
What the law hard-codes (ED can’t redefine)
Who’s in the program cohort. Only Title IV–aided completers from the academic year 4 years before the determination, who are not enrolled and are working. ED can define “working,” but cannot add non-completers or include non-aided students for pass/fail. Congress.gov
How many years to pass. A program fails if its cohort’s median earnings are below the comparator in at least 2 of the 3 most recent years. ED can’t change the “2-of-3” structure. Congress.gov
Comparators & geography (undergrad). Compare to the median earnings of 25–34-year-old working adults with a high school diploma in the state of the institution—or U.S. national if <50% of students reside in-state. Data source must be the Census Bureau. No MSA/ZIP/COLI swaps here. Congress.gov
Comparators (graduate). Compare to 25–34-year-old working adults with a bachelor’s degree, using the lowest of: state-wide general, same-field-in-state, or same-field national (with a different “<50% in-state” fallback). “Field of study” is as determined by the Secretary (e.g., 2-digit CIP), but the “lowest-of” rule itself is statutory. Congress.gov
Small cohorts. If a program’s cohort is <30, ED must aggregate additional years, then aggregate equivalent-length programs to reach ≥30. The 30 minimum is in law. Congress.gov
Student notices, appeals, and re-entry. ED must provide an appeals process, require student notices after 1 miss in the 3-year “covered period,” and allow re-entry after ≥2 years via an ED-set process. (ED designs the process; the existence/timing is statutory.) Congress.gov
Effective timing. Institutions must assure compliance beginning July 1, 2026 (the law’s start date). ED can phase process details but can’t move the statutory date. Congress.gov
Common asks from comments — what’s realistic? A) Doable now via regulation (high confidence)
Define “working” carefully. ED can set earnings/wage-match rules (e.g., minimum positive earnings, treatment of self-employment/1099, unemployment insurance gaps) consistent with the statute’s “working” requirement. Congress.gov
Choose “field of study” granularity for grad comparators. Law lets ED decide the taxonomy (e.g., 2-digit vs 4-digit CIP) for graduate-level “same field” comparators. This is a big lever for fairness/noise. Congress.gov
Spell out program mapping & cohort construction. ED can define what counts as a “program” (credential level, CIP, length), how to handle cross-listed CIPs, teach-outs, and transfers—so long as it stays within the statute’s program-level frame. Congress.gov
Design the appeals & re-entry processes. ED decides evidence allowed (e.g., data-quality issues, misclassification), timelines, and documentation standards. Congress.gov
Operational safeguards for small programs. While the ≥30 minimum is fixed, ED can define “equivalent length,” the order of aggregation, suppression rules, and privacy protections. Congress.gov
Student disclosure specifics. ED can set format, content, and delivery of at-risk notices, and require standardized language. Congress.gov
Measurement cadence & data handling. ED can lock the reference years, how it stitches multi-year results, and which federal data it uses for program cohort earnings (the statute dictates Census for the comparator, not for the cohort measure).
More to Come
This is a massive change in a citizen’s ability to track public comments during rulemaking. I even wonder of ED has the ability yet to summarize all this information as it comes in.
Note, however, that this is partial data.
The public comment period ends on Thursday (Aug 28th).
Only 557 of the 815 submitted comments (to date) have been posted.
Most people add comments at the end of the comment period.
Therefore, there will be hundreds, if not a few thousand, additional comments posted by next week.
But it will be easy for me to rerun this process and provide a summary of the full set of public comments.
On EdTech+ premium subscribers can access the Python script and the combined CSV file to enable their own analysis.