As higher ed grapples with the complexities and expenses of admissions, the allure of adopting AI grows.
Admissions is a complicated and expensive function of higher education. As a result, some institutions are considering adopting artificial intelligence.
Although AI offers the promise of streamlining certain admissions tasks, it also brings numerous challenges that institutions are trying to wrap their arms around.
Admissions offices have been using AI for some time in the form of chatbots to answer general questions. However, the technology is expanding into more complex corners of the admissions process, explained Dr. Diane Gayeski, a professor of strategic communication at Ithaca College.
A September 2023 survey by Intelligent.com of 399 education professionals with in-depth knowledge of their schools’ admissions processes found that half of the schools were already using AI in their admissions offices. Additionally, 82% said their offices would be using it by 2024. Efficiency was cited as the top reason for AI’s use.
Dr. Stephen Keller, senior director of undergraduate admissions at Gonzaga University, believes AI is here to stay. For that reason, he said, admissions professionals will start to utilize it as a tool.
“In much the same way that a calculator speeds up the mundane calculations of mathematics to allow time for higher level thinking, AI has the very real potential to speed up administrative tasks in offices of admissions,” Keller said. “Gonzaga’s Office of Admission is always looking for ways to spend less time with processes so that we might spend more time with people.”
Selective institutions continue to struggle with the volume of applications they receive and the complexities of holistic reviews, said Jerome Lucido, professor of clinical practice at the USC Rossier School of Education and scholar-in-resident at the USC Center for Enrollment Research, Policy and Practice.
“They seek ways to streamline the process, without removing the human element, through the use of AI,” Lucido explained.
Dr. Royel Johnson, associate professor at USC Rossier School of Education and director of the National Assessment of Collegiate Campus Climates in the USC Race and Equity Center, agreed with Lucido on why institutions find the use of AI attractive.
“AI significantly enhances efficiency in admissions by rapidly processing applications and automating tasks,” he said.
As a result of streamlining the admissions process, institutions can reap some real economic rewards, said Don Hossler, provost and Sonneborn Professor Emeritus in the School of Education at Indiana University Bloomington.
“The goal of admissions processes should be to admit students that fit the institution and will be academically successful,” said Hossler. “To the extent AI can help institutions streamline that process so that they spend less money on the enrollment management part of institutions, that’s a good thing. Universities don’t exist to have the most expensive enrollment management functions.”
According to Hossler, the enrollment management function is now one of the largest and most expensive administrative areas for institutions.
“Places where we can save money and still make good decisions, we should absolutely be trying to do as much of that as possible,” he said.
Hossler believes that institutions may reduce the number of admissions staffers they use to review applications if AI tools make the process more efficient.
“The director of admissions at UC Berkeley talked about the amount of money and the number of part-time people they’d hired to go over all the applications once they decided to switch to a more holistic review,” he explained. “It was a pretty staggering number of people and a fairly expensive initiative as well. I think it’s possible in some areas that it will reduce expenses.”
Some services, like Student Select AI, are being used to solve situations like the one at UC Berkeley that Hossler mentioned.
“Artificial intelligence is helping admissions teams manage the numbers game of applications and admissions to better support a holistic applicant review process and more inclusive environment,” the data science company indicated in an article.
According to the Intelligent survey, the use of AI to review transcripts is widespread among institutions that use AI in admissions, totaling 71%.
“That’s something that can be time-consuming and the systems are proving to be fairly robust and accurate in doing that kind of thing,” Gayeski said. “It can flag things in an accurate way that even human readers who get very tired after reading hundreds of essays may not.”
In addition to freeing admissions offices from the time-consuming work, she explained, using AI to review transcripts also gives applicants a quicker answer on how many of their credits will transfer.
Humans, she added, will still likely be used to double-check the AI-generated reviews.
Experts expect AI to be used for more sophisticated tasks, such as admissions modeling based on student characteristics and the review of supporting materials.
“As AI’s role in admissions evolves, its applications will likely expand beyond chatbots and predictive analytics to include more sophisticated analysis of essays and personal statements,” Johnson said.
Gayeski offered the example of using AI to screen GPAs as one of the ways it can help institutions model student characteristics.
“Things like a group of students that there’s no way they would be considered. Here’s another group who look like they absolutely would be considered. And then there’s a big pile in the middle. Making some of those initial kinds of categorization can help,” she said. “It’s a good first step. And in some cases, it can highlight things readers might want to pay attention to.”
She anticipates schools with large numbers of applicants to general programs will probably be more eager to adopt AI than smaller institutions and ones with niche programs will. For example, she said AI would not be helpful for arts-based programs. However, some platforms use AI to view video interviews and analyze the transcripts.
“They’re interesting in that maybe they can provide a quick summary that helps people have an advanced organizer as to what they’re looking at in a video or maybe sort out people who can’t address certain things at all,” she said.
Lucido added that the technology can also be used to review essays.
The Intelligent survey found that, among respondents whose schools currently use AI in admissions, 87% say that AI ‘sometimes’ (43%) or ‘always’ (44%) makes final decisions about whether to admit applicants or not.
Gayeski said she is not aware of a specific institution using AI to make final decisions. However, she cautions against this particular use, except in obvious cases.
“If one is applying to a very competitive program and a student puts forward their transcripts and they have mostly Ds and Fs, I guess I wouldn’t mind if AI did make the final decision that they’re not going to be accepted because I think that’s pretty obvious human beings would too,” she said.
The technology could also decide if programs mandate specific requirements and applicants fail to meet those requirements, she added.
Johnson said, “Using AI to make final admissions decisions is concerning due to its inability to capture the holistic essence of an applicant’s potential and character.”
Although Hossler can imagine AI being used by selective institutions to winnow the field, he doubts they will use it to select students for admission. Additionally, he flatly believes that humans should make final admissions decisions.
“Human beings have had a significant role in programming the AI that they’re currently using, and should anybody actually let AI make the decision, I think they ought to be totally transparent about that, which I doubt would happen,” he said. “I doubt behind the veil of admissions, anybody’s going to want to admit that that’s what they’re doing if, in fact, they are.”
Gayeski said, that if the standards used to train AI models aren’t clear, there will be a lot of variability, which ultimately will lead to questions about the entire admissions process.
“The use of AI requires us to be more transparent and consistent in our decision-making, which isn’t a bad thing. Because you have to program it so carefully, you really have to articulate what it is you’re looking for,” she said. “For some people, it’s such a black box. They aren’t sure what schools or individual departments are looking for. There’s a nuance there that many colleges and universities have different criteria for different programs.”
She pointed to an example of a prospective student applying to an acting program. Their acceptance will rely more heavily on their audition and not as much on whether they received an A in their physics class.
“There are programs that are looking for things that applicants might not realize. They’re just so focused on how many clubs they’re involved in, being president or captain of a team or GPA,” she explained. “In fact, a lot of departments look for different kinds of things, like if they have experience in that field and know what they’re getting into.”
If institutions use AI, being able to articulate what they are looking for will help level the playing field and allow applicants to highlight the things that matter in new ways, Gayeski said.
Institutions, Gayeski added, can make their standards clear by issuing statements on the admissions pages of their websites. “Here are some things we find are correlated with success in our students and you might want to highlight [them] in your application,” she said.
Two-thirds of the Intelligent survey respondents reported that they are concerned about the ethics of AI.
AI, as with all processes, can advantage some and disadvantage others due to built-in biases, Lucido explained.
Johnson said, “Technological innovations like AI are baked with biases that can reproduce inequities. It’s people who design and inform the algorithms, curate the data and make the decisions that shape these systems.”
Lucido cautioned that AI models used in admissions must be built so that the programmers are cognizant of historical and emerging biases.
“Its use must be meticulously researched to ensure that inequitable consequences are not a result of its use,” he said.
He added that nuance surrounding an application could be overlooked if institutions use AI improperly.
“Institutions must ensure that fairness, equity, and individualism are maintained in their file review,” he said. “The nuances are important. So, too, are the underlying contexts of student backgrounds.”
Johnson also stressed that, although AI can streamline the admissions process, it lacks the human capacity to appreciate the depth of personal essays fully, potentially overlooking the unique, intangible qualities that make applicants stand out.
“This limitation underscores the importance of maintaining human oversight in the review process,” he said. “AI is only as just as the equitable decisions that inform its design.”
Admissions directors, Hossler stressed, need to think very carefully about what’s going into the creation of the AI tools they’re using.
“Think about where the information is coming from,” he said. “If the information is flawed, if the information is biased, that means that the AI itself is going to be biased. In looking at issues of fairness and equity, institutions have an obligation to be very careful in terms of looking at their data and some of the decisions they’ve made in the past to make sure that the AI isn’t contaminated.”
AI, Gayeski noted, now calls into question whether the admissions process in general is biased.
“Admission is based on who has done well at our university in the past. What does a successful student look like? That can be quite biased given the long history of who got into higher ed and who did well for any variety of reasons,” said Gayeski. “I think we have to look at our criteria overall.”
Institutions are cognizant of the need to consider ethics related to the use of AI in admissions, Gayeski added.
“There are so many concerns about diversity, equity and inclusion in the admissions process now anyhow,” she said. “They’re aware that AI can have a bias.”
Gayeski isn’t sure whether institutions should make it clear to prospective students that AI is being used to review their applications because they aren’t transparent now about how the admissions process works.
“I don’t think they know a lot of the categorization is done by readers who are part-time, fairly low-paid employees. They’re just scanning for very specific things using a rubric, not unlike what AI would do,” she explained. “You’re not getting, for example, a psychology professor reading psychology applicants. You’re just getting some reader who is looking at everything from physics to physical therapy to zoology. It’s a very mechanical process right now.”
However, knowing that their applications are being reviewed by a real person is important to prospective students.
“There are always going to be things that can’t be trained or understood by an AI model,” said Gayeski. “At least a sympathetic human will hopefully take the time to dig a little or flag something or ask more questions.”
Hossler agreed that prospective students and parents will want to know if AI is being used. “Institutions do have some obligations,” he said. “If any school that emphasized their personalized approach to admissions and that everybody gets a personal review — if at some point any of those schools change and AI does those things — then I think institutions have an obligation to let students and parents know what they’re doing.”
Stay current in digital strategy, brand amplification, design thinking and more.
STEM-designated MBAs offer international students extended work opportunities while benefiting higher ed institutions facing enrollment pressures.
Faced with increasing competition and limited resources, colleges are turning to AI technology to streamline admissions and effectively manage student recruitment.
As higher ed grapples with the complexities and expenses of admissions, the allure of adopting AI grows.
STEM-designated MBAs offer international students extended work opportunities while benefiting higher ed institutions facing enrollment pressures.
Faced with increasing competition and limited resources, colleges are turning to AI technology to streamline admissions and effectively manage student recruitment.
Only time will tell if the one-year confusion for the ‘simplified’ federal aid form will yield long-term benefits.
