Higher Ed Uncovering New Ways to Apply Data and AI Across Campus

• By Dian Schaffhauser
• 03/14/19

The use of data analytics and artificial intelligence, the inevitability of data breaches and six other topical subjects made up the latest "State of Higher Education" report from consultancy firm Grant Thornton. The company, which provides audit, tax and advisory services to 200 public and private institutions, each year covers the trends and issues it sees emerging out of the sector to help higher ed leaders understand how to manage their schools through the changes.

In the area of data analytics, audit services experts Claire Esten and Natalie Wood pointed to "program costing" as one of the nontraditional areas where the use of data is helping to contain cost. As an example, data analysis could be applied to future program development by examining classroom use and margins at both the "major and course level," which could lead to changes in the mix. Data could also play a role, they wrote, in predicting fraud and identifying noncompliance. Examples included data analysis to find out if student athletes were being given preferential grading treatment by faculty; or determining if some form of fraud was occurring tied to student financial aid by examining patterns in online applications, withdrawal dates and the timing of student refunds and loan applications.

The use of AI could also be applied to business operations in similar ways. As Audit Partner Rick Wentzel and Advisory Services Manager Raisa Reyes stated, institutions are using AI for "robotic process automation," data analysis work and better "customer service through chatbots and automated workflows." Robotic process automation, as they explained, "is AI that takes information and manipulates in a repeatable or predictive manner to automate repetitive, time-consuming work." For instance, some schools have begun using it to review applicants' documents and identify missing information, compare the contents of transcripts to admission criteria, and run algorithms to do a "first pass through" on college entrance essays (which can check for plagiarism or multiple submissions). In another example on the academic side, algorithms are helping instructors at Cornell analyze wrong answers on exams to identify how students might have arrived at their incorrect conclusions, helping faculty to understand where they need to "revisit topics" or "modify their teaching style" for improved student learning.

The coverage of "inevitable" data breaches profiled two universities facing the same dilemma: phishing e-mails encouraging recipients that use self-service HR platforms to click on a link that would allow them to log into their accounts. Those credentials would be scooped up, allowing scammers to go into the account and redirect direct deposits to another account (and change e-mail addresses that would alert the victim to the change). One college with "an adequately documented incident response plan" "sprang into action" to prevent funds from being transferred and mobilized a response team to "investigate and contain the nefarious activity" within a few hours. Another school without such a plan took four days to identify victims, establish the cause of the problems and contain and eliminate the breach. Full recovery took another couple of weeks, and the institution lost about $340,000 as a result. Not surprisingly, Grant Thornton Advisor Hassan Khan's advice focused on building a "robust incident response plan" with delineated roles, tabletop exercises, wide communication and regular updates.

One theme was woven through every topic in the report: the application of new technologies and practices to "effect substantial operational change," as the head of the higher education practice, Mark Oster, noted in his introduction. But both of those also require "innovative thinking" for "successfully moving into the future."

The report is available with registration on the Grant Thornton website.