Some promising applications of artificial intelligence are being used in teaching and learning, student success, and accessibility. However, there is still anxiety and concern about data governance and algorithmic bias that may prevent widespread use on higher ed campuses.
Recently, EDUCAUSE sent out a Quick Poll to determine how widely AI is being used by colleges and universities for instruction, institutional tasks, and student success and support tasks. Poll results revealed that AI is being used for monitoring student behavior in exams and detecting plagiarism. AI is least used in institutional tasks. Many respondents were unaware of AI use, perhaps indicating that it is so embedded in their programs that it is invisible.
The primary use of AI in higher ed is to preserve academic integrity through plagiarism software (60%) and proctoring apps (42%). These uses have been around for quite a while but increased during the remote learning phase of the pandemic. Apparently, both of these uses of AI are being examined over concerns about violating student privacy and producing false positives. The majority of respondents said that they are not using AI and have no plans for it in instructional tasks, such as providing student feedback, tutoring assessments, and grading.
Where AI is being used is supporting student success. 36% of respondents noted that chatbots and digital assistants are in use somewhere on campus. 17% are planning, piloting, or are in initial stages of use. Chatbots are used by admissions, student affairs, career services and other student support services. Student success tools may grow the use of AI. Currently, it’s being used for:
- Identifying students who are at risk academically – 22%
- Sending early academic warnings—16%
- Planning, piloting, initial use—14%
The author of the poll summary, D. Christopher Brooks, Senior Analyst at Tambellini Group, speculated that because the reporting appears low, that respondents were not considering their use of analytics, thinking that it’s not considered AI. In fact, analytics is a subset of AI. According to respondents, AI is not being used for curriculum planning, financial aid, admissions decisions, or development/fundraising. Other uses of AI include:
- Analysis of student evaluations of teaching
- Instructional planning
- Social media analysis
- Support desk services
- Attendance on campus
Again, there is the possibility that low awareness of AI use in higher education is an ability to see it as it is baked into the apps and tools that administrators, faculty, and students use.
Challenges and Promising Practices
Two-thirds of the respondents noted that they have institutional deficiencies that prevent adoption and maintenance. Three-quarters responded that ineffective data management and integration (72%), insufficient technical expertise (78%), financial issues (67%), immature data governance (66%), and insufficient leadership support (56%) are the leading challenges. There are also concerns about ethics of AI use (68%) and algorithmic bias (67%).
Promising practices include a range of current usage in the following:
- Chatbots for informational and tech support, HR benefits questions, parking questions, service desk questions, and student tutoring
- Research applications, conducting systematic reviews and meta-analyses, and data science research
- Library services
- Recruiting students
- Providing individual instructional pathways, assessment feedback and adaptive learning software
- Proctoring and plagiarism detection
- Student engagement support and nudging
- Detection of network attacks
- Recommender systems
While there appears to be no danger of AI being used to replace instructors, it is likely that AI is more prevalent than administrators understand as it is embedded in digital programs and tools. However, AI chatbots are handy when they “liberate staff” from repeatedly answering the same questions, or reduce errors. AI can also be useful when automation is required to accelerate digital processes such as content recommendation engines. It is clear from the Quick Poll results that colleges and universities are not yet convinced that the benefits of AI outweigh the concerns.