Integrating AI into higher education

Artificial Intelligence (AI) can be controversial when we think about it as an educational tool. Yet, there is a plethora of opportunities where colleges and universities can effectively integrate AI into their teaching and learning.

I had the opportunity and great pleasure to write about AI in higher education as part of my work reporting on EdTech for Fierce Education for several years before the publication came to an end. It was rather sad for me to say goodbye to one of the publications I enjoyed the most. I worked with a fantastic Head of Content and a lovely team of writers. The Head of Content crowned me their "expert in AI," a title I earned after 18 years of investigating and reporting on AI and related technologies and applications.

During my years with Fierce Education, I had the pleasure of interviewing many faculty members, researchers, policy makers, and industry leaders on EdTech topics. The conversations were insightful. It was nice to see how educators could adopt or were adopting X or Y technology in their classroom. Of course, AI in general and GenAI such as ChatGPT were on the menu.

On one occasion, the insight came from Michael Vaughn, Education and Adoption Specialist at Open LMS, working on a Generative AI use policy.

"AI is great for dramatically speeding up time-consuming tasks. If you were building a deck and needed to drill in screws to hold the lumber in place, you'd likely use an electric drill instead of a hand screwdriver. Both produce the same result, but the electric drill is faster and more efficient than doing the task by hand. AI works in a similar way with tasks that are traditionally cumbersome or data-intensive."

Vaughn believes that there is great potential for AI to positively impact higher education. He told me that if used thoughtfully and responsibly, AI be used in the following scenarios:

• Admissions: An AI tool could be trained to assess student recruitment data. This tool could analyse potential applicants, flagging those that have a higher likelihood of success, and suggesting financial aid packages that would likely lead to an application.

• Instructors can (currently) use Generative AI tools like ChatGPT to help them dramatically reduce the time it takes to build course assessments. Say, an instructor is using the same assessment in two different courses and is worried the students in Section A will share the answers with the students in Section B. Tools like ChatGPT can help the instructor rapidly remix a second version of each question to test the same concepts, but with different question framing and responses.

• Even student support services can benefit from AI-powered tools and platforms. The Career Services Center can help students write persona prompts that help ChatGPT temporarily adopt the role of interviewer for a job the student is applying for. This gives the student a credible, no-stakes dry run of the interview beforehand, helping the student craft and refine the responses to interview questions.

Of course, there are still challenges and gaps related to AI's adoption in the classroom. According to Vaughn, if you asked most faculty what the biggest challenge with AI is, they would probably say "students cheating."

"I think it goes further beyond academic integrity," he says. "If I came to you and said I had an AI-powered platform that would analyse hundreds of data points about your students, and identify which ones were at risk of dropping out of school because of academic performance, that would sound like a great tool to have. But what data would the AI need to access to work?"

"Again, AI is great at analysing incredibly large datasets to find patterns that are otherwise difficult to spot," says Vaughn. "What student data does that tool need to work? Will we give it access to course grades? Grades on individual assessments? Attendance data? What about the number of times they visit the campus wellness centre, or if they are using counseling services? Do we use card swipe data at dining halls and campus events to track which students are forming peer groups? There's a very fine line between support and surveillance with tools like this, and it's not always clearly defined," he told me.

Vaughn tells that on the academic side faculty are already overwhelmed with ever increasing demands on their time and the new unavoidable technology they need to learn to use.

"Though some technological innovations come and go, I do believe AI is here to stay in the world of education. Even if a school were to ban its use and implement specialised detection software to catch academic dishonesty, it won't be enough to dissuade students from finding and using these tools."

Vaughn encourages faculty to think of AI-powered tools like ChatGPT the same way that math teachers think of calculators.

"Students have access to an incredibly powerful tool, but it really only works well when you know how to use it. I can hand anyone a TI-83 graphing calculator. That doesn't mean they will know how to generate a graph of an algebraic equation without the proper instruction and guidance from a subject matter expert. If we expect students and faculty to learn to competently use word processing software, email, and learning management systems (LMSs) it's not unreasonable to add AI-powered tools like ChatGPT into that collection of technological tools."

On the topic of what faculty and higher education leadership can do to ensure the proper use of AI in the future, Vaughn strongly recommends trying to catch up as quickly as possible.

"Instagram is one of the largest social media platforms in history, and it reached 100 million active monthly users in two and a half years. ChatGPT reached 100 million active monthly users in two months," he said.

"Generative AI tools like ChatGPT are not going away, and they continue to improve at a very rapid pace. Failing to develop policies and guidelines around AI now will almost certainly lead to a campus-wide AI ban later. And if you reach the point of banning AI on campus, you've deprived your students of the opportunity to develop what will likely be an extremely marketable skill in the future. That may even put your students at a disadvantage compared to competitors," Vaughn said.

Vaughn recommends higher education leadership to consider developing a school policy regarding AI needs. Such policy should cover at least the following areas:

• Definitions of what constitutes AI

• Guidelines for responsible usage of AI-powered tools, services, and platforms (including any tools the school or the students use)

• Relevant and illustrative examples of both appropriate and inappropriate uses of AI

• How the school defines transparency related to the use of AI

Pretty much the point of teaching how to use ChatGPT and other technologies to all students is to create impactful and engaging learning experiences for today's learners.

***

About the Author: Susan Fourtané is a freelance Science and Technology Journalist, a professional writer with 18 years experience writing and reporting for global media and industry publications. She's a member of the ABSW, WFSJ, Society of Authors, and London Press Club.