This story is not about ChatGPT. Most admissions offices are terrified that some (many?) students will find ways to use the new technology to submit essays that are not their own.
But that doesn’t mean that AI hasn’t entered admissions offices—by their invitation.
Some admissions offices are starting to use AI, for instance, to review transcripts, which are a key part of a college application.
College officials, association leaders and the companies involved all insist that systems are safe and that there is plenty of (human) checking of the work of machines. And they all insist that no one is being admitted entirely through AI.
But a few colleges are starting to use AI technology, and to talk about it.
One of them is Maryville University, a small private college in Missouri (if you are counting only traditional undergraduates on campus) that has 6,500 students enrolled online.
Phil Komarny, the chief innovation officer at Maryville, said the university recently contracted to use Sia (a tool from the company OneOrigin) to begin reviewing applicants’ transcripts in September.
Komarny stressed that the university’s use of Sia isn’t “just about the automation of reading the transcript.” When the AI reviews the transcripts, it enables admissions officers to spend their time on other things, he said.
It’s about a change in philosophy at Maryville for IT functions from “control” to “community.” He added, “We are a collective obsessed by data.”
Banshan Syiem is vice president for sales and marketing at OneOrigin. He said his company has about 22 higher education customers using Sia. Generally, they pay $500 to $3,000 a month, depending on the number of transcripts being reviewed.
“We’ve eliminated the work,” he said.
“We read the transcripts from high schools and colleges,” he said. “It knows the content is there.”
The college transcripts that are being reviewed are for transfer admissions, he said.
“Many colleges have criteria for admissions,” Syiem said. “If you are a student at Maricopa Community Colleges and you want to transfer to Arizona State University, we can help.”
“Instead of going through all the applications,” an admissions officer “can get the information instantly.”
He stressed that his company is trying to help admissions officers spend more time making decisions and is not trying to make the decisions about whether to admit a given applicant.
Other colleges are experimenting with AI in admissions.
Stephen W. Harmon, executive director of the Center for 21st Century Universities at the Georgia Institute of Technology, said Georgia Tech Is “not yet using AI in admissions, but we are experimenting with it.”
One project is to try to replicate admissions decisions using machine learning techniques. “We are using one of our large online master of science programs as a test case, and we are currently at about a 93 percent match with our admissions advisers’ decisions,” Harmon said.
What is the appeal? “The volume of applicants continues to increase at Georgia Tech, so anything we can do to improve their workflow is helpful. I don’t see us relying solely on AI for admissions maybe ever, but it could become a useful tool in the process, “ he said.
What the Experts Say
The experts that help admissions officials are cautious on AI, but see its value.
“One of the most significant ways AI is impacting the college admissions process is by automating the evaluation of applications,” said a blog post from North Shore College Consulting. “Many institutions receive tens of thousands of applications each year, making it a Herculean task for human evaluators to sort through and assess each one. This is where AI comes into play—software can quickly and accurately analyze large swathes of data, enabling universities to streamline their admissions process and spend more time focusing on the finer aspects of student evaluation.”
Others say it may be too early to declare any consensus.
“We are hearing a lot of buzz about the potential implications of AI in and around the college admission process,” said David Hawkins, chief education and policy officer of the National Association for College Admission Counseling, via email. “Right now, there isn’t anything approaching consensus about where AI should or shouldn’t be used, though there is a common thread running through many conversations that I believe will form the basis for future best practices.”
Hawkins added, “Specifically, while AI can process, summarize and potentially create efficiencies for tasks that heretofore have been labor-intensive, it is only as good as the directions you give it and the quality control you exercise on the back end. It’s therefore clear that anyone utilizing AI will need to be careful about constructing the AI’s instructions and monitoring its output.”
And he called it “unlikely at this point that institutions would entrust high stakes decisions to AI processes or that such a move would be considered best practice in the future, though there is a great deal still to be discovered about the potential and limits of this new technology.”
Melanie Gottlieb, executive director of the American Association of Collegiate Registrars and Admissions Officers, said, “I have not yet heard of using AI to make admissions decisions. We are hearing about using it in the field in very functional ways like making transfer recommendations, reviewing transcripts, etc. Using AI to manage rote functions more efficiently seems to be in the future. The admissions decision is not necessarily a rote function unless you are talking about an access institution.”
Gottlieb added, “For competitive institutions, I think there may be concerns in the application of AI to a process that the public already finds opaque and challenging. I might also have concerns regarding inadvertently hard-coding of bias into AI models. These conversations are certainly coming, but we aren’t having them quite yet.”