Behind closed doors, in an increasingly competitive higher education marketplace, admissions offices across the country are using artificial intelligence to identify prospective students, flag potentially AI-generated essays and sift through mountains of transcripts.
“This is perfunctory, soul-sucking work that ideally you could not have humans doing,” said Rick Clark, vice president of enrollment management at Georgia Tech.
The Atlanta university began using AI this fall to review grades and test scores for nearly 68,000 applicants competing for about 4,100 spots.
“A lot of our students might go to two high schools during their high school career, or they might dual enroll at another college, so it’s not just 68,000 transcripts; sometimes it’s 150,000 transcripts coming in,” Clark explained. “That means a human counting these rigorous courses. It also means a human finding sometimes a weighted and an unweighted GPA.”
Colleges across the spectrum — from public flagships to selective private institutions — are using AI to review essays, sort applicants by test scores and GPA, and, in some cases, make full admissions recommendations. The technology is also being deployed to help determine financial aid packages, raising new questions about how algorithms evaluate students and their families as well as concerns about privacy and transparency, fairness and equity.
On Monday, a former Justice Department official said the Trump administration is exploring using AI tools to investigate alleged discrimination in admissions. Speaking at Harvard Law School, former Civil Rights division chief Harmeet K. Dhillon said while the division has relied on statisticians to analyze admissions data, staff are now considering AI “within the bounds of government policy.”
Student advocates are sounding the alarm.
“I think students should know exactly how an admissions office is using AI so that they know how it’s scoring their application and different biases that may come with that,” said Aaron Ament, president of Student Defense, a legal and advocacy nonprofit that works to advance economic mobility for students seeking higher education.
Earlier this year, Ament’s group filed public records requests and found no public university had an AI governance policy specifically addressing admissions. Student Defense has since developed a “Student AI Bill of Rights,” which calls on colleges to disclose if, when and how the technology is used to evaluate applicants.
Ament attributes the growing use of automation to budget cuts, declining state support and falling tuition revenue as the pool of potential students who can afford what colleges charge shrinks.
“They’re seeing where they can cut costs, and then use AI not just to replace a lot of human functions and save money in the administration of college — recruitment and admissions — but then also try to get as much tuition dollars as they can out of students,” claims Ament.
Industry insiders defend the practice.
“As large language models in particular have continued to grow, colleges have used them in a lot of different ways,” said Joseph Miller, an independent educational consultant based in Houston.
In the past, Miller has worked both in high school counseling and college admissions offices. Today he leads the National Association of College Admissions Counselors Admissions Practices Committee. He says AI-usage varies widely. Some colleges have dropped these tools entirely and hired more human readers instead. But others — including Virginia Tech — have been using them for years.
“I think for the most part, there is a lot more transparency and we’ve called for that just so everybody’s on the same page,” he said.
Leveraging the latest algorithms and AI agents to estimate what families can afford — and how much they might be willing to pay — may seem like a straightforward, innocuous way to set tuition prices. But Spelman College President Ayanna Howard worries AI systems trained on past admissions decisions could disadvantage students who don’t fit established, historical patterns.
“If you do not look like the students that have been admitted in prior cycles and have been successful, the AI might not tag you as being of high rank,” said Howard, who recently wrote “Rebooting the Machines,” a book arguing that humans should shape technology to reflect humanity rather than allow it to replace it.
Howard says Spelman, a historically Black college, does not use AI directly in admissions, but it does rely on outside data tools to project enrollment and retention.
Howard says humans must remain the final decision-makers and colleges must address potential bias before it becomes embedded in their systems.
“At some point, I’m pretty sure we are going to evolve to the fact that AI is just going to do it all, and you’ll have a human somewhere, but really late in the cycle,” she predicted.
Her advice to students navigating the world of college admissions in the age of AI? Ask colleges whether they use AI in admissions. Write essays in your own words, and don’t try to game an algorithm or mold yourself into someone you think a college wants to see. Instead, show up as your whole self.
At Georgia Tech, Clark agrees with that advice. He says a trained chatbot now handles routine, frequently asked admissions questions, freeing up his staff to focus on what he calls “higher-value work.”
“If humans don’t have to spend as much of their time answering emails or answering calls, they’re freed up for the things we want them to do, which is build relationships with students — stay on the road longer in recruitment,” he said. “That’s where we’re bullish on some of the innovations here.”
On average, Clark estimates, the new transcript-sorting technology saves three to five minutes per application. But he stresses the fate of prospective Georgia Tech students isn’t being determined by AI.
“No, 100%, AI is not making admission decisions — not making financial aid decisions,” he insisted in an interview, shaking his head. “No.”
For now, those human decisions remain in human hands.