Higher education in America is a massive industry, with 20 million students enrolled in college today. Artificial intelligence tools, like those used to predict what stocks will increase in value, often rely on predictive algorithms for decision making. Predictive analytics uses historical data to predict future events.

Predictive algorithms depend on loads of data to make their predictions. They tell us what movie we might like to watch next and if we might get Alzheimer’s in a few years. In the context of a university, they can use historical data about students to streamline admissions and retain students.

For instance, algorithms can speed up the selection process. When applications and test scores are received, they can sort them into three groups: accepted, declined, and maybe. The next step would be to generate the appropriate emails and letters to go out to students.

For a university of 6,000 students, admissions officers receive anywhere from 20,000 to 40,000 applications. That number represents mountains of applicant essays, transcripts, test scores and letters of recommendation to pour through and evaluate. In the future, predictive algorithms that thrive on large data sets could conceivably detect patterns that could establish the right fit for a university. AI would be able to learn which behaviors predict consistency and commitment, for instance.

Should an algorithm have the last say on whether an applicant is admitted or not?

Any AI predictions will be based purely on facts, and that can never paint a complete picture of a person. AI could help to narrow down the number of applicants that would be a good fit for a tertiary institution, but it probably should never make the ultimate decision about whether an applicant should be admitted or not. Universities want to admit applicants that are a good fit for them and have the greatest potential of success at their institution. This helps them with student retention in the long run.

Preventing “summer melts”

A matter that is related to admissions is the so-called “summer melt”. That happens when applicants that have been admitted to a university or college, don’t turn up to start their studies. Georgia State University in Atlanta has begun using an AI chatbot system called Pounce, to reduce these incidents.

During the summer of 2016, Georgia State created a list of more than 2000 questions and answers for freshmen. The questions pertain to issues like financial aid, courses, majors, housing, etc. In fact, many applicants don’t understand how admissions processes operate. The university then collaborated with the conversational AI-company AdmitHub to deliver those answers to students on text-based platforms they could access anytime via their smartphones.

By the time fall classes started, more than 200,000 questions by freshmen had been answered. The university credits the Pounce system for helping it to reduce the summer melt that year by 22%. The answers themselves were written by humans, not by AI.

These are just a few uses of artificial intelligence that can help tertiary institutions perform better. AI can help institutions to enroll applicants that are a good fit for the institution, are likely to remain for the duration of their course and stand a good chance to succeed in their studies.