AI in College Admissions Statistics 2026: Sourced Data

The numbers worth quoting, with publisher, sample and date

For US colleges, two surveys do most of the work in 2026: EAB on how students use AI in the college search, and Kaplan on how admissions offices set rules for applicants. Both were opened at the primary source. Two UK studies add context and are labeled as such. None of the numbers comes from Skolbot.

A note on method: no paid keyword-volume tool was available for this piece. The topic was scoped from public search results and primary sources.

Prospective students: nearly half now use AI in the college search

EAB's survey of more than 5,000 high school students, fielded in October and November 2025 and released on February 25, 2026, found that 46% use AI tools in their college search. In spring 2025 the figure was 26%. Eighteen percent had taken a college off their list because of what an AI search surfaced.

That second number is the one an enrollment team should read twice. It means the first impression of your institution may be formed by a summary you did not write. It also means accuracy of your public pages (costs, deadlines, programs, aid) now affects who applies.

For the student-side picture beyond the US, HEPI reports 95% of UK undergraduates using AI in some form. These are enrolled students in another country, so use them as context only.

Admissions offices: most have not written the rules yet

Kaplan polled admissions officers at selective US colleges by email in July and August 2025 and published results in November. On using generative AI to write the admissions essay, 2% allow it, 30% ban it and 68% have no policy. The ban share rose from 25% in 2024. On brainstorming, 27% allow it (18% in 2024); on feedback for a draft the applicant wrote, 21% allow it (15%).

Fifty percent of respondents view generative AI in admissions essays unfavorably, 26% are neutral and 14% favorably. The release mentions "more than 200" officers in the text and 220 colleges in a footnote; we cite "about 220 colleges" and flag the discrepancy.

We did not find an open, primary survey that measures how many US colleges use AI to read or score applications. Figures that circulate on this come from a 2023 poll mixing K-12 and college respondents, so we have left them out.

The research stage in the UK offers a reality check

The UCAS research, summarized on Wonkhe by UCAS's Senior Insight Lead, surveyed 4,485 people in November 2025. Forty-three percent said a university website is their first stop; 13% would start with a chatbot. Forty-eight percent would drop a university after a bad open day, versus 8% after a negative AI review.

It is a UK sample, with a different application system, so it does not predict US behavior. It does suggest that campus experience and your own site still carry weight alongside AI answers. Compare it with EAB's trend, do not merge the two.

Legal context for US admissions teams

Student data rules in the US are layered. FERPA protects education records of students; applicants who never enroll are generally outside its scope, but once a student enrolls, admissions records can become education records. State privacy laws, the FTC's consumer-protection authority and your institution's own commitments to applicants also apply, and requirements vary by state.

We found no comprehensive federal AI statute governing admissions tools. Some state laws address automated decision-making, and the rules are changing, so confirm with counsel which apply to your applicants. The EU AI Act lists systems that determine admission to educational institutions as high-risk (Annex III), with obligations applying from December 2, 2027 per the European Commission. It can reach a US college recruiting applicants in the EU.

On quality, regional accreditors (such as HLC, MSCHE, SACSCOC and WASC) govern accreditation. They are not AI regulators, but your admissions claims must remain accurate under their standards. The Common App and the Coalition platforms set their own application integrity rules for students.

What these figures do not show

They do not show that a chatbot raises yield or applications: none of the sources measures conversion. They are not comparable with each other: high school students, admissions officers and UK undergraduates answered different questions in different months. And they say little about community colleges, graduate programs or adult learners.

For published conversion references, see our article on school website conversion rate benchmarks.

How to vet an AI statistic before you quote it

Ask four things. Who ran the survey, and who paid for it? How large was the sample and whom does it represent: applicants, parents, officers? When was it fielded? What exactly was asked? Many numbers circulating on social media fail on the first question because they cite no primary source. Every figure in the table above passes, within the scope stated.

What an enrollment team can do with this

• Audit what an AI says about your college. Ask the main assistants about your costs, deadlines and programs, and fix the public pages that feed wrong answers.
• Publish your AI policy for applicants. Most colleges in the Kaplan sample have not.
• Keep people in the admit decision. Use AI for questions, routing and scheduling.
• Track your own inquiry data: top questions, response time, tour sign-ups.

Further reading: our guides to AI agents for student recruitment, the definition of an AI admissions agent and the AI chatbot student recruitment guide.