How Artificial Intelligence is Reshaping Higher Education




Like every corner of our economy, higher education is integrating artificial intelligence into its processes and programs in ways you might expect—and in some you might not.

Harnessing the power of AI and machine learning seems like a perfect answer to the massive administrative burden carried by most academic institutions. The thinking goes: If we can let AI do the bureaucratic grind work, it will free up faculty and staff to do work only humans can do.

With the application of any new technology, it’s critical to look at early-use cases for validation.

Here are some of the ways that AI is currently being used in Higher Ed:

Application Screening

In one application of AI recently reported on by the Wall Street Journal, colleges and universities have been using machine learning in an attempt to address the problem schools nationwide are facing. Massive application numbers and steadily lowering “yields” (the number of students actually accepted).

By applying machine learning to prospective student engagement analytics, universities reported greater efficiency and greater accuracy in determining the viability of students, honing in on the so-called “demonstrated interest” of applicants.

Advanced Tutor Apps

One of the biggest problems any learning institution faces is the limited time–in the classroom, for individual feedback, and due to personal responsibilities. One on one attention between professors and students is essentially impossible in today’s current university environment.

However, additional interaction and feedback is massively helpful for most students. Artificial intelligence is looking to address this with tutoring applications.

Applications like these can provide instant feedback on quizzes, questions, or reports so that students are engaging more frequently with their studies. This allows professors and students more time to connect socially and develop critical thinking skills.

Instructional Data

When instructional designers build curriculums, they rely on student and teacher feedback, grades, and pass rates. By leveraging more software and artificial intelligence, there are swaths of additional data that can be useful for improving courses.

Data surrounding teaching effectiveness, student weakness, and learning disability information can all be used to shape better courses.

Student Support

For students, universities are implementing AI-driven systems that can offer support in some personalized ways.

Chatbots – Recently, Arizona State University developed a messaging platform for students and faculty that includes an AI chatbot named Devi. Devi can help students manage their courses, tackle technical issues, and even offer career advice. Devi’s not alone. Many universities are offering chat-based guidance for students looking for advice on courses and careers. These applications use vast data sets of similar students and success rates.

Financial Aid – Other areas of student support utilizing AI include financial aid processes—including just-in-time aid that can automatically identify imminent financial needs and help students procure aid— and creating early warning networks that use data from academic and operational sources to help identify students at risk of failing or other substantive issues. These red-flag systems are going beyond simple GPA monitoring. They recognize data points like–if a student starts skipping meals in the cafeteria. Being able to identify these patterns can help universities re-engage these students before it’s too late.

Improving Mental Health

University of Denver Assistant Professor Anthony Fulginiti is exploring how social network analyses and other AI-informed processes can help identify and prevent potential suicide risks. It also helps during follow-up procedures involving suicide survivorship and prevention hotline engagement.

Also at DU, the Ritchie School of Engineering and Computer Science is growing its faculty for a new University-wide initiative around AI for the Public Good. This initiative focuses on critical policy, ethical, legal and business issues in addition to basic algorithmic, machine learning, data science, and computer vision AI and data science work.

Artificial Intelligence in higher education is not without its controversy. The Wall Street Journal example stated earlier has its opponents who cite privacy concerns of applicants and, as with any application of machine learning to human issues, there is also the problem of bias either from an algorithmic or implicit/programmed source. If the algorithm factors in data with an unintentional bias (for example, how zip codes can indicate race or ethnicity) or if the creators of the program bring with them implicit bias, that then gets instantiated into the machine learning. Since AI results tend to carry the supposition of unassailable truth and impartiality—”numbers don’t lie”—these very real biases can go undetected or unchallenged.

While we’re not quite at the point of handing over the keys of higher education to the bots, there certainly are interesting and valuable applications of AI happening in academia today, with plenty more opportunities on the horizon.