A group of college accreditors is backing the use of artificial intelligence to reduce credit loss during transfer, which is a major barrier to completion for many of the 43 million people across the nation with some college credit but no degree.

“Current approaches frequently result in delays in students receiving necessary information about how their credits will transfer, the need to retake courses, and other negative consequences for students,” said a statement the Council of Regional Accrediting Commissions released Monday. “Technological advances, such as AI, can help institutions improve this process.”

Although the statement from CRAC, which represents seven accrediting agencies that oversee 3,000 total degree-granting institutions, isn’t a mandate, the council hopes it will send a message to colleges and universities that leveraging AI to expand course equivalencies doesn’t conflict with accreditation standards.

“There’s a myth that accreditors don’t allow for certain kinds of considerations during credit evaluation,” said Heather Perfetti, president of the Middle States Commission on Higher Education and chair of CRAC. “Accreditors do actually promote innovation and encourage institutions to reconsider the ways they are managing transfer.”

Credit Loss

While the majority of community college students say their goal is to transfer to a four-year university, fewer than half of those who transfer earn a degree within six years. That’s in part because transferring credits to a new college is an inconsistent, time-consuming and decentralized process that often increases the cost of and time to degree completion.

In 2024, four in ten adult Americans reported trying to transfer credit, and 58 percent of them reported some credit loss, according to a survey from the nonprofit research group Public Agenda; 20 percent of those who tried to transfer credit reported having to repeat a class they already took because their credits didn’t transfer. Having to repeat a course caused 13 percent of respondents to run out of financial aid. For 16 percent, the process of transferring credits was so difficult that they gave up entirely on pursuing a degree or credential.

Faculty and administrators say manually evaluating courses for transfer is too labor-intensive. At the same time, research shows that decisions about which transfer credits to accept are often guided by inconsistent, arbitrary guidelines and bias; more white respondents (76 percent) than Black (66 percent) and Latino respondents (64 percent) said their new college accepted all or most of the credits they tried to transfer, according to the Public Agenda survey.

‘Cultural Shift’

But AI—which can analyze large data sets related to course descriptions, enrollment and learning outcomes—can help streamline the learning-evaluation process and reduce credit loss.

“Faculty want to know if what is taught on a student’s transcript is comparable to what they are teaching in terms of the learning objectives,” said Zachary Pardos, an associate professor of education at the University of California, Berkeley, who has studied AI and course evaluation. “In my research and other reports that have come out, most faculty are looking for about a 70 percent overlap, and AI is well positioned to find out if the courses have that overlap.”

In 2024, Pardos partnered with the American Association of Community Colleges, the Association of Public and Land-grant Universities and the consulting firm Sova to launch the AI Transfer and Articulation Infrastructure Network. This year, a pilot cohort of 59 institutions is using the platform CourseWise, which engages AI to analyze course equivalencies between participating colleges and identify new or expanded matches that count toward degree completion.

AI, he added, has the potential to usher in a “cultural shift” away from faculty erring on the side of not accepting a course for transfer.

“CRAC’s statement gives leadership at institutions the right moment to make a decision to do something innovative, whether that’s exploring the approach we’re experimenting with right now or other innovative approaches,” Pardos said. “The accreditors aren’t saying it’s time for a free-for-all, where we give credit for everything. It’s that AI is at a sufficient level of accuracy that you can use a trust-but-verify approach.”

Pardos’s work also helped to inform CRAC’s statement, which implores institutions to “commit to a default in learning evaluation that credits are applied to program completion unless there is evidence that the required learning outcomes are not met,” combined with considering AI for learning evaluation. “Decision-making should not be based upon anecdotes, assumptions about quality, locations where earned, or an unexamined history of ‘how things have always been done.’”

John Fink, a senior research associate and program lead at the Community College Research Center at Columbia University’s Teachers College, said he’s hopeful that CRAC’s push to get institutions to explore AI may help “flip the system from the default of ‘credits do not transfer unless there is evidence that they should’ to the default of ‘courses that meet certain equivalency criteria using predictive algorithms are deemed transferable unless there is evidence to the contrary.’”

However, “the devil will be in the details of how AI is being used to identify which courses are equivalent and how reliable this is when implemented on a large scale,” he added. “But if the field can figure this out, it would certainly speed up transfer credit evaluations and help students while also building in checks from academic departments as needed.”

CRAC’s statement suggests that AI and other technological innovations may be used to:

• Reduce credit loss for students by analyzing existing course equivalencies and identifying new or expanded matches so more classes count toward degree completion;
• Provide students with critical information about degree-applicable credit in a timely manner;
• Reduce the administrative burden of learning evaluation
• Free up faculty and staff time to focus on teaching, mentoring and guidance rather than paperwork

“This statement is about encouraging institutions to examine deeply their current learning evaluation practices. That could be examining how AI could be leveraged to support some of the challenges with processing transfer,” said Perfetti, the CRAC chair. “Certainly, there is controversy around AI and the ways that it can and should be used. But it’s clear that AI can play a solid role in transfer credit and get at some of those barriers we know exist with transfer credits: the slow nature of processing, the inequity of decisions and the loss of credit without good reason or rationale.”

The goal of CRAC isn’t to take humans out of the course evaluation loop entirely, but to enhance their decision-making about which transfer credits to accept, said Stephen Pruitt, president of the Southern Association of Colleges and Schools.

“Evaluating course credit, transfer and content gets so decentralized that it can be interpreted many different ways,” he said. “What AI should be able to do is make that a little more seamless and aid in that evaluation so that it doesn’t depend on which professor gets to decide if the course material is appropriate or aligned. There’s actually something that puts that initial stake in the ground that provides humans with more uniformity in how they approach that work.”

‘Force’ of Accreditors

The CRAC statement is also a sign from accreditors that they encourage embracing AI to advance student success, at a time when many higher education institutions have been slow to develop comprehensive AI guidelines.

“We’ve been silent on it, and we can’t afford to remain silent. As long as we stay silent, others make assumptions about what they can or can’t do with that technology,” Pruitt said. “CRAC just simply releasing a statement saying, ‘This is a good thing and you should embrace it’ isn’t going to make anything different or better. It’s going to be institutions stepping up to the plate and deciding how they’re going to use it and [make it] work for them.”

But the CRAC statement may help put some momentum behind the idea of using AI to improve course evaluation, said George Railey Jr., vice chancellor for academic success for the Alamo Colleges District and a member of the Beyond Transfer Policy Advisory Board.

“Educational entities are slow to change, but things are moving so rapidly. Enrollments and budgets are declining, and colleges need a way to bring more students to their campuses,” he said. “[CRAC’s statement] puts the force of the accrediting commission behind efforts to move institutions into the present and perhaps the future with AI to advance student completion and success.”

(This article has been updated to better reflect the results of Public Agenda’s survey.)