Everyone says students use AI for everything. Ask the students.

What employers report

The employer-side numbers are consistent and alarming. Willo's 2026 Hiring Trends report found 77% of hiring teams regularly encounter AI-generated applications, up from 53% in early 2024. ZipRecruiter's 2026 survey found 48% of employers say AI has increased the number of applications per role. LinkedIn told the New York Times it processes roughly 11,000 applications a minute, up 45% year over year. Robert Half found in March 2026 that 67% of U.S. HR leaders say reviewing AI-generated applications has slowed their hiring, with one in five reporting delays of more than two weeks.

From inside a recruiting inbox, it looks like everyone is using AI to apply everywhere.

What students report

NACE's 2025 Student Survey ran April 1 to May 30, 2025, with 13,684 students responding, including 1,479 graduating seniors at the bachelor's level. Thirty-three percent of those seniors said they used AI in their job search. The ones who did used it mostly for cover letters (64.8%), interview preparation (63.8%) and resumes (61.6%).

The 2026 Student Survey, with more than 17,000 students across 258 institutions, found use had grown but still fewer than half of graduating seniors used AI in their search. Half of them, 50.5%, said they weren't building AI skills for the future. Thirty-one percent said AI skills would be of little or no importance to their career. Students who didn't use it gave plain reasons: they'd rather do it themselves, they didn't need it, they never considered it. Some cited environmental cost. Some said it wasn't accurate or helpful.

From inside a senior's job search, AI is a tool some people use for cover letters.

Why both are true

Employers see applications, not applicants. A minority of candidates sending a large volume accounts for a large share of what arrives. NACE's own data shows how this works: the average senior submitted about 30 applications, up from 16 or 17 two years earlier. Students who used an AI search engine, roughly 15% of respondents, averaged about 60. Run the arithmetic and that 15% produces close to a third of all applications. An employer that sees a third of its inbox as AI-driven is seeing a real thing. It's just not the median student.

Employer surveys also measure belief. ZipRecruiter found only 24% of employers say they can almost always tell when AI was used. "Regularly encounter AI-generated applications" is a perception, and perception rises with volume regardless of cause.

And the employer's pool isn't made of new graduates. Indeed Hiring Lab reported that in May 2026, 62% of applications from job seekers with six to nine years of experience went to entry-level roles. The inbox is full of experienced workers, career changers and people applying from overseas. A survey of employers about "applicants" isn't a survey about your seniors.

The part that should worry a career center

The finding that matters isn't the low usage. It's the gap between what students think and what employers now expect.

Half of graduating seniors say they aren't building AI skills. Meanwhile NACE's Job Outlook 2026 Spring Update found more than a third of entry-level jobs now require some AI competency, nearly triple the share in the fall survey, when only 10.5% of entry-level job descriptions mentioned AI. That shift happened inside a single academic year.

And students aren't being taught. NACE's fall 2025 quick poll found fewer than a third of students reported having any guidance on using AI in the job search, and about one in five had any instruction in using AI as an assistive tool. Career centers themselves use it heavily: 86% report using AI as an assistive tool when working with students in 2026, up from 20% in 2023. Only 40% had run AI workshops for their own staff.

So the center uses it, the employer expects it, and the student mostly doesn't get taught it.

What this means for career services

Don't design your advising around the doom-loop image. The student blasting 200 AI-written applications exists, and NACE's data suggests that student is also less likely to end up in a job related to their major. But that's a minority, and the intervention for them is different from the intervention for everyone else.

For the minority producing volume, the work is redirection. Fewer applications, better targeted, with the AI doing tailoring rather than multiplication.

For the majority, the work is instruction. NACE found students who'd heard the term "skills-based hiring" were more confident in their skills and more likely to list competencies on their resumes. Familiarity changes behavior. The same is likely true for AI: students who've been shown how to use it to verify, tailor and prepare will use it differently from students who've only heard it's cheating.

Close the gap between staff practice and student instruction. If 86% of centers use AI with students and 20% of students have been taught to use it, the knowledge is in the building and isn't being transferred.

What would change this reading

If the 2027 Student Survey showed a clear majority of seniors using AI in their search, the "minority producing volume" explanation would need revising. If employer detection improved substantially, the perception argument would weaken. If NACE's entry-level AI-competency figure fell back toward the fall 2025 level, the urgency on instruction would drop. None of those has happened.