It's AI all the way down: Colleges are using AI to screen applications

Colleges are using AI to read college essays, analyze transcripts, interview students about research projects and check for fraud, report Jaweed Kaleem and Jocelyn Gecker in the Los Angeles Times. Schools say AI tools improve efficiency and are less prone to bias.

Of course, it raises the prospect of AI tools evaluating AI-created applications, so the chosen can use AI to earn a credential whose value will be determined by AI-using employers. As the old joke goes: It's turtles all the way down.

This fall, early-admissions applicants to CalTech who submitted research projects were interviewed on video "by an artificial intelligence-powered voice that peppered them with questions about their papers and experiments, akin to a dissertation defense," report Kaleem and Gecker. "The video-recorded exchanges were then reviewed by humans — faculty and admissions officers — who also evaluated test scores, transcripts, personal statements."

Virginia Tech's AI-powered essay reader will make it possible to inform applicants much sooner.

“Humans get tired; some days are better than others. The AI does not get tired. It doesn’t get grumpy. It doesn’t have a bad day. The AI is consistent,” says Juan Espinoza, vice provost for enrollment management at Virginia Tech.

Instead of two humans scoring each essay, Virginia Tech is using one human and AI. If the two disagree, a second human will step in.

"Like many colleges, Virginia Tech has seen a huge increase in applications since making SATs optional," they write. Reading all those essays takes time, even with 200 readers.

I can think of a way to solve the problem without AI. Require applicants to submit an SAT or ACT score. Fewer will apply. Don't bother reading essays from applicants that score below the minimum -- unless they're star athletes or the offspring of major donors. Or, to make it even simpler, don't require essays. The odds that they're not the student's own work are too high.

California's community colleges are using AI to identify phony "cash-and-dash" students who sign up for federal and state aid, take the money and run. Online classes have made this kind of fraud even easier. New AI services can often spoke the fakes.