Four in Ten Students Already Ask AI About Colleges. What Are They Learning About Yours?

For the first time in a decade of tracking how prospective students research their college options, a traditional Google search is no longer the default starting point for a significant share of students. AI is.

A Decade of Listening, and a Year That Changed Everything

OHO has spent more than ten years studying how prospective students research college. Across 150+ surveys, 23,000+ respondents, and 700+ interviews and focus groups, we've watched the research process evolve through the rise of social media, mobile-first search, and now generative AI.

The 2026 survey — fielded in March 2026 with 530 respondents, including 327 prospective and current undergraduates ages 17–22 — captures the most significant behavioral shift we've seen in that entire decade. It’s not a wholesale transformation of how all students search. But for a large and rapidly growing segment, the college search has been reorganized from the ground up.

The Numbers That Should Get Your Attention

Among students ages 17–22 in our sample:

  • 68% have used AI chatbots generally
  • 73% of those actively researching college (or recently enrolled) use chatbots routinely
  • 40% of that actively-researching segment have used AI specifically for their college search

In other words, roughly four in ten prospective students who are actively researching college are already doing so in ways that look fundamentally different from the pre-AI model. And among those students:

  • AI tools and traditional search engines are essentially tied as the first source they turn to (36% start with AI, 32% with Google)
  • 65% rate AI-generated information as accurate — up from 51% just two years ago
  • 74% say AI helped them discover schools they hadn't previously considered
  • 50% say AI helped them remove schools from their consideration set
  • Nearly three-quarters have received answers about a college without ever visiting its website

These aren't fringe behaviors. They're the defining characteristics of a research process that is already reshaping consideration sets — at institutions that, in many cases, have no idea it's happening.

AI Isn't Replacing Search. It's Replacing the Counselor.

The most important finding in the 2026 research isn't how much students are using AI. It's how they're using it.

AI isn't functioning as a passive information source. It's acting as a decision-guiding partner. Students are asking AI to recommend schools, narrow their lists, compare finalists, estimate their odds of admission, and project their post-graduation salaries by region and specialization. They're asking questions like "Based on what you know about me, will I fit in here?" — and treating the answers as meaningful input.

That's not search behavior. That's advising behavior. And it's happening at scale, on demand, in a conversational format that many students find more accessible than a scheduled meeting with a counselor.

The Credibility Illusion

Here's the finding that should change how you think about your institutional content: students trust AI responses largely because of the presence of source citations — not because they actually click through to verify them.

In our ethnographic research, one student put it plainly: "I trust ChatGPT a lot because it's giving me valid sources. You know it probably is reliable if it gives you a source." When we asked whether he ever clicked through to check, he said: "I don't always check the sources, But I know they are there, so I trust them for the most part."

This is what we're calling the credibility illusion. The source list functions less as an invitation to verify and more as a visual cue that the answer is trustworthy. Which means that for a growing share of prospective students, the impression your institution makes is being formed before they ever touch your website — and if the AI summary is incomplete, outdated, or subtly wrong, you may never know, and the student may never question it.

Meet Nico and Tia

The research goes deep on two student portraits drawn from our ethnographic work: Nico, a first-generation engineering student in Baltimore for whom AI functions as a risk-reduction engine, and Tia, a composite of socially-oriented students for whom AI serves as a fit-and-identity engine.

Watching Nico navigate ChatGPT in real time was a revelation. He didn’t open Google first. He asked ChatGPT for engineering programs near Baltimore, drilled down on a school his counselor had mentioned, asked about salary outcomes for engineers in Maryland — and got a regionally adjusted projection in seconds that no college website could have produced. When he finally visited the school's site, he ignored the institutional chatbot in the corner. He'd already gotten his answers from a tool he trusted more.

Tia's pattern is different but structurally identical. Where Nico arrives at AI with a concrete question, Tia arrives with a feeling — "I want a medium-sized school with a creative vibe somewhere on the East Coast" — and uses AI to translate that feeling into a list she can evaluate. Then she goes straight to Instagram to check whether students themselves confirm what AI said.
Different motivations. Same structural role for AI: it's the environment in which the decision starts to organize itself.

What This Means for You

  1. Your content has a new audience: machines. What AI says about your institution depends on what it can find and parse from your website and other public sources. Clear, structured, plain-language content on your highest-stakes pages — admissions, cost, programs, outcomes — is no longer just a UX best practice. It's an enrollment imperative.
  2. Your website has a new job — but hasn't lost its old one. For AI-using students arriving to validate, your site needs to deliver specific facts in one or two clicks. For fit-focused students and the substantial segment not using AI at all, it still needs to tell your story with warmth and authenticity. Both audiences must be served.
  3. Audit what AI says about you — now, and on an ongoing basis. Open ChatGPT and Gemini. Ask the questions your prospective students are asking. Document the gaps and errors. Then fix them. This needs to become an ongoing institutional function, not a one-time project. 
  4. Don't abandon traditional channels. Roughly one in four students ages 17–22 doesn't use AI chatbots at all, and a meaningful additional share chooses not to use AI for their college search specifically. These students still rely on search engines, websites, counselors, and peer networks. The right move isn't to pivot away from traditional digital investment, it's to add a new, parallel strategy.
Jim Dalgish, Director, UX Research and Strategy

Jim Dalglish

Director, UX Research and Strategy Published: