AI has moved into admissions, advising, tutoring, grading and research faster than any technology higher education has absorbed. The guardrails have not followed. Between 2019 and 2026, state legislators filed 115 bills on AI in K-16 education and passed 14. Only one of those laws, Tennessee’s 2024 S 1711, tells colleges anything about how AI may be used. Washington has offered funding guidance and a fight over state authority, but no safeguards.
So colleges and universities are governing AI themselves. This post looks at what the legislative record shows, and at what it leaves campus leaders to build on their own.
Campuses are not waiting
Institutions are building AI governance faster than legislatures are. In EDUCAUSE’s 2025 AI Landscape Study, 39% of respondents said their institution had an AI-related acceptable-use policy, up from 23% a year earlier, and 57% called AI a strategic priority.
However, only 22% reported an institution-wide AI strategy; 55% said it was rolling out unevenly across departments. Just 9% said their cybersecurity and privacy policies were adequate for AI risks. A 2026 Jenzabar survey of nearly 200 institutions found 26% with a formal institution-wide AI policy and 35% with none or unsure, while 48% of staff wanted clear acceptable-use rules.
In other words, expansion is real, and so is self-regulation. What is missing is are common expectations for student data, academic integrity, disclosure, vendor accountability and human review of high-stakes decisions.
Washington offers money and preemption, not safeguards
The federal posture leans toward adoption. In July 2025 the U.S. Department of Education issued a Dear Colleague letter telling schools that AI projects can be paid for with existing federal funds, and proposed a grant priority covering AI in K-12 and higher education. That is guidance on spending, not rules on use.
In December 2025 an executive order, “Ensuring a National Policy Framework for Artificial Intelligence,” went further in the other direction. It created a Justice Department task force to challenge state AI laws and tied some federal funding to states not enforcing laws the administration views as conflicting. It set no binding federal AI rules of its own. For colleges, the message is that state safeguards may be contested while federal ones have not been written.
A few days ago I came across The National Conference of State Legislatures (NCSL) and its repository of every state AI bill, 2019 through 2026. I discovered 2,053 bill records across the 50 states and DC related to AI.
NCSL’s “Education Use” and “Education/Training” tags are too broad on their own. They sweep in state budgets, workforce bills, and even resolutions declaring “Carolina Day.” Another challenge was multiple carryover bills, so I merged them and kept the latest status.
I kept bills tagged for education that name schools, students, educators, districts, colleges, or universities, plus untagged bills whose titles mention schools. I also dropped general budget bills, ceremonial resolutions, and included bills where AI use in education is what the bill governs, funds, or studies. That leaves 115 core K-16 AI bills in 34 jurisdictions. I coded each one by policy type (governance, study, literacy, funding, school safety) and by level (K-12, postsecondary, or both). The full coded file is available for replication.
Finding 1: The surge came late and fast
For five years, AI in education barely registered: 11 core bills from 2019 through 2023. Then 89 arrived in the 2024 and 2025 sessions alone.
The education spike rode a wider wave. All state AI bills rose from 75 in 2019 to 474 in 2024 and 1,136 in 2025. Yet K-16 bills stayed a small slice of that wave, about 7% of AI bills in 2024 and 5% in 2025. Legislators were far busier with deepfakes, elections, health care and private-sector AI than with classrooms.
Finding 2: Most bills never move forward
Of the 115 core bills, 14 were enacted or adopted (12%), 53 failed or were vetoed (46%), and 48 were still pending when NCSL’s data was exported (42%). Among bills with a final outcome, roughly one in five passed.
The type of bill matters. Governance bills, the ones that require policies, guidance or limits on AI use, were the most common (40) and produced the most laws (6).
Study bills and task forces were the least productive: 26 filed, one adopted. Several states, Mississippi and Missouri among them, filed the same task-force bill three years running without passing it.
The outright bans did worst. Texas filed bills to prohibit AI in classroom instruction (H 2400, S 382) and to restrict AI scoring of state tests (H 4390, H 5282); all failed. New York’s 2026 bills to bar most classroom AI before high school remain in committee. However, we all know that the governor of New York banned AI from K-12 classrooms.
Finding 3: The states that talk most are not the states that govern
Two groups emerge. Introducers file heavily and pass little: Illinois, Maryland and New York filed 30 bills between them. Implementers file less and enact more: Utah, Tennessee, Ohio, Virginia and Nevada passed working policy with a handful of bills each.
The 14 enacted bills in the data, newest first:
| Year | State | Bill | What it does | Level |
|---|---|---|---|---|
| 2026 | Utah | H 273 | State Board must write model policies on classroom AI and technology use | K-12 |
| 2026 | Virginia | S 394 | AI safety guidance plus an AI Innovation in Education pilot | K-12 |
| 2025 | Illinois | S 1920 | State Board must issue statewide AI guidance for districts and educators | K-12 |
| 2025 | Ohio | H 96 | Budget bill requiring a model AI-use policy for schools | K-12 |
| 2025 | Nevada | A 406 | Restricts certain AI uses in public schools (mental-health context) | K-12 |
| 2025 | Louisiana | HR 320 | Resolution asking K-12 and higher-ed boards to promote AI education and PD | Both |
| 2025 | Maryland | H 782 | Study of weapon-detection technology for schools | K-12 |
| 2025 | Delaware | H 16 | Adds a high school student to the state AI Commission | K-12 |
| 2024 | Tennessee | S 1711 | Public colleges must adopt AI rules; districts and charters must adopt AI policies | Both |
| 2024 | Florida | H 1361 | Grants to districts to implement AI for students and teachers, with use requirements | K-12 |
| 2024 | Connecticut | H 5524 | Funds an AI education-tool pilot with educator PD | K-12 |
| 2024 | Nebraska | L 1284 | Research grants for AI writing support for students with dyslexia | K-12 |
| 2021 | Mississippi | H 633 | K-12 computer-science curriculum that includes AI | K-12 |
| 2020 | Utah | S 96 | Deep-technology talent initiative in higher education | Postsecondary |
Tennessee’s 2024 law stands out. It was the first in this data to require both school districts and public college governing boards to set AI-use rules for students, teachers and faculty, and it did so in one short bill.
Two cautions on the scoreboard. First, California shows zero enactments, but SB 1288 (2024) was recorded as “to governor” as of October 2026; the California Department of Education now runs the AI working group that bill created. Second, 2026 enacted far more than this data shows: trackers list Idaho SB 1227, Maryland SB 720, Oklahoma SB 1734, Alabama HB 329 and Utah HB 218, among others.
Finding 4: Seventeen states never started the conversation
Seventeen states filed no core K-16 AI bill in the NCSL data from 2019 through 2026. They fall into two tiers.
| Tier | States | What the data shows |
|---|---|---|
| Fully silent | Alaska, Idaho, Kansas, New Hampshire, South Dakota | No education-related AI bill of any kind. South Dakota filed 2 AI bills in total across eight years. |
| Peripheral only | Alabama, Arkansas, Georgia, Iowa, Michigan, New Mexico, North Dakota, Oregon, Rhode Island, South Carolina, Washington, Wyoming | AI appears only in budget lines, ceremonial resolutions or computer-science mandates, never as the subject of school or campus governance. |
The peripheral tier is not empty of money. Michigan funded Michigan Virtual to promote responsible AI use in schools, Iowa funded AI degree programs at Iowa State, and Maryland routed budget dollars to Morgan State’s equitable-AI center. Those are investments, not rules.
The 2026 sessions are already shrinking this list. Outside trackers report that Idaho enacted a statewide K-12 generative AI framework (SB 1227), Alabama tied AI instruction to a graduation requirement (HB 329), and Oregon and Washington passed AI tool-safety measures. As of the sources I checked, Alaska, Kansas, New Hampshire and South Dakota remain without a K-16 AI bill.
Finding 5: For colleges, the legislative record is nearly blank
Legislators are governing AI in K-12, not in higher education. Of the 115 core bills, 88 address K-12 only, 9 span both levels, and 17 address postsecondary only.
The postsecondary bills that did appear are mostly reports and capacity-building: Illinois asking its Board of Higher Education to report on AI (H 5399, H 2655), California proposing AI workgroups at CSU Long Beach and across UC, CSU and the community colleges (S 1235, SCR 82), and AI hubs or institutes in North Carolina and Hawaii. One Illinois bill (H 1859) would bar community-college courses from replacing the primary instructor with AI.
Only one enacted law sets AI-use rules for colleges: Tennessee’s S 1711. Private and proprietary institutions are almost invisible. Only two bills even mention private higher education, both as subjects of a report or a training partnership, and neither passed.
For leaders at private and for-profit institutions, the silence is complete. There is no state mandate to meet and no state framework to lean on, so every safeguard on campus is one the institution chose to build.
What this means: expansion without a floor
The legislative record and the campus surveys tell the same story from two sides. AI use is expanding fast. Public safeguards are not. The gap is being filled by institutions, unevenly.
- Illinois, Ohio, Utah and Virginia told a state board to write guidance; Tennessee told each district and college to adopt its own rules. The substance of AI governance is being written either way locally.
- 26 study and task-force bills produced one adoption. Hearings and commissions signal concern without committing anyone to a rule.
- Without a statutory floor, a campus that gets AI wrong in admissions, grading or student data faces the consequences alone, and a campus that gets it right has no common standard to point to.
In the absence of state and federal frameworks, what AI safeguards do postsecondary institutions adopt, and who drives them?
Limitations and sources
NCSL’s 2024 file still lists some bills as “to governor,” and its 2025 file listed Illinois S 1920 as pending although it became law in August 2025 (corrected here). Many “pending” bills in two-year sessions may yet pass; many in one-year sessions are effectively dead.
The 2026 data is a partial, education-only pull of 17 bills. FutureEd counts 87 AI-in-education bills in 28 states for 2026, and MultiState counted 134 in 31 states by March. The 2026 bar and the 2026 enactment count here understate the year.
The K-16 filter and policy-type codes are my judgment calls. Borderline cases, such as school weapon-detection bills and computer-science mandates, are documented in the coded file so others can recode them.
Sources
- National Conference of State Legislatures, Artificial Intelligence Legislation Database, 2019–2026 exports (author’s spreadsheet)
- FutureEd, Legislative Tracker: 2026 State AI in Education Bills (updated Oct. 6, 2026)
- ExcelinEd, State K-12 AI Policy in 2026: Milestones, Momentum and Missing Links (May 26, 2026)
- MultiState, How States Are Regulating AI in Education This Legislative Session (April 2026)
- California Department of Education, AI in Education Workgroup (SB 1288)
- infoDOCKET summary of the 2025 EDUCAUSE AI Landscape Study (Feb. 15, 2025)
- EdScoop, Higher education warming to AI, EDUCAUSE survey finds (2025)
- Jenzabar, 2026 AI Pulse: As Optimism Grows, So Does the Need for Policy (June 10, 2026)
- BABL AI, U.S. Department of Education Releases AI Guidance, Proposes New Grant Priority (July 2025)
- DLA Piper, New Executive Order Aims to Preempt State AI Regulation (Dec. 2025)