Artificial Intelligence (AI) and machine learning are the buzzwords of the 21st century. From music and movies, to healthcare and banking, this cross-industry evolution has become a silent helper, making our lives easier without us even realizing. But while AI is dominating progress in other industries, we're only just starting to hear about AI in higher education.
Back in 2018, we published a three-part series delving into what was then “emerging technology”, giving you a rundown on the basics of machine learning, its potential to transform higher education, and how schools can begin integrating it into their own processes. Then, AI in higher education was an interesting idea. Today, it’s being put into practice.
Driven by the pandemic and a growing demand for digital solutions, AI has become a key tool for institutions looking to keep current and prospective students engaged with their academics and campus community. In admissions, evaluation of personality and soft skills are helping teams identify and establish early interest with top candidates. In classrooms as well as admissions teams, chat bots are supporting students and applicants by acting as an ever-present guidance counselor or TA. And in academic success departments, data analysis is helping to identify students’ pathways to success, while simultaneously raising flags for at-risk students to receive targeted help and support in both social and academic contexts.
We’ve done a deep dive to bring you the best and brightest from AI in higher education. Discover how you can start transforming your student data into intelligent actions at every step of the student lifecycle.
5 ways you can leverage AI in higher education
- Flag applicants early in the admissions process
- Create a contact point for incoming students
- Enable 24/7 chat bot support in online (and in-person) classrooms
- Flag at-risk students who may need support
- Identify pathways for student success
1. Early applicant flagging in the admissions process
If your goal is to reduce workload on administrative teams, one of your focuses should be enrollment and admissions. Each year, admissions teams sift through hundreds if not thousands of applicants to find the best fits for their programs. And while we know that the best recruiting tool is to show early interest in an applicant, it can be difficult for small teams to balance a thorough reviewing process with a quick turn-around time. Implementing an AI screening tool can help.
Today, AI is being used to pre-screen incoming applicants and flag high-potential candidates for early review. By training the software to identify a program's defined success markers (be it grades, volunteer experience, etc.), the tool helps you find and fast-track top applicants.
If you’re using video assessments in your admissions process (high five!), then AI can also help you pre-screen for soft skills. AI software can evaluate applicant videos for speech, text, verbal, and non-verbal cues. By comparing its findings to its database of catalogued video content and cross-referencing the Big Five personality traits model, the tool generates a personality assessment for each applicant. Your team can then identify the applicants with the character traits and soft skills required to succeed in your program, and they can start reviewing those applications right away.
Discover how assessing personality traits can predict success in medical students
While it’s important to always have a human hand guiding admissions decisions, AI has come a long way in supporting that process. By expediting preliminary tasks, AI allows for teams to spend more time giving each applicant due focus.
2. Creating a contact point for incoming students
The “summer melt” is a big problem for schools, as students, especially first generation students, start to feel overwhelmed by the impending financial and mental commitment of college.
One of the best ways for schools to avoid summer melt is to maintain their connection with students. Whether it’s getting them excited about campus, sending them information on semester prep, or helping them sort out the financial side of things, frequent contact has been shown to reduce the melt. Access to a knowledgeable person who can give them guidance and encouragement during this overwhelming time can make a huge difference. Unfortunately, most schools aren’t able to provide that kind of 24/7 support to every student. However, AI has provided an answer.
in higher education in real life
The University of Murcia in Spain recently brought on an AI chatbot to help the admissions and enrollment team keep connected with incoming students. The chatbot took question and answer scenarios from the school’s email records and online forums, and used them to provide incoming students with information about the campus, programs, financial aid structures, and more. If the bot couldn’t find an answer, it would generate an email connecting the student to someone in the admissions department.
By the end of the first testing phase, the school was able to answer the students’ questions correctly 91% of the time. This totalled over 38,000 questions which the admissions team would have normally dealt with.
The school found that access to the chatbot increased motivation and reduced summer melt, as students were able to get an immediate response to any question without feeling the pressure or potential embarrassment of sending an email.
And the bonus? The school was able to pull the chatbot’s records and use the data to determine what areas of focus were important for their incoming students. Were they more concerned about their courses? Or about campus living? Did they have questions about financial aid? Or grading schemes? This not only helped the school identify areas where there seemed to be a lot of confusion, but it also helped them re-evaluate what they should highlight in their brochures for the next class.
3. Chat bot support in online (and in-person) classrooms
Information fatigue is a real problem for higher ed students. Whether it’s course material, administrative info, or extracurricular announcements, it’s easy for things to fall through the cracks.
This leads to an ever-relatable story of professors becoming frustrated by a multitude of student questions on something that was clearly outlined in the syllabus. But Professor David Kellerman at the University of New South Wales (UNSW) may have found the ultimate solution using AI.
Prof. Kellerman worked with Microsoft to build a chatbot which could recognize and answer students’ questions by pulling from a wide database of previously asked questions, as well as lecture materials, textbooks, digitally published guides and the school’s own question forums. If the bot didn’t understand the question or couldn’t find the appropriate answer, it was able to forward the student's request to a relevant TA.
This 24/7 AI support not only freed professors and assistants from answering repetitive questions, it also drove a sharp increase in student course satisfaction with the class garnering a 99% approval rating.
UNSW and Microsoft have open-sourced this software. You can check it out here.
4. Flagging at-risk students who may need support
Students today are juggling more than ever before. Between a full course load, extracurriculars, social engagements, and often multiple minimum-wage jobs, it’s no wonder that many burn out before completing their degree.
With the overall six-year graduation rate sitting at around 60% for students studying in the US, it’s clear that schools need to invest in systems to better assist their at-risk students. AI can be that support.
Predictive analytics generated by AI can act as early warning systems for students who are at risk of failing classes, missing payments, or are struggling with mental health issues. By analyzing various data points, including academic standing, peer and professor feedback, and on-campus activity, AI can generate an overview of a student’s behaviour. Instead of simply relying on grades to determine well-being, this creates the opportunity to develop a more holistic understanding of the student’s overall “wellness”.
In addition to being more timely, the AI software can flag students when they hit certain triggers or fall into identified risk patterns, providing up-to-the-minute assessments instead of the traditional mid- and post-semester review.
5. Identifying pathways for student success
Doubt surrounding the selection of majors and academic goals accounts for one of the top 5 reasons why students drop out of college or university.
For undergraduate students in particular, post-secondary can represent a maze of different programs, focuses, and course options. For many, generating a schedule can, in itself, be an overwhelming task. Here, guidance counsellors and academic advisors play a key role in helping current and prospective students understand requirements and map out an educational plan that is best suited to their needs. However, as these teams are often strapped for time, they are unable to reach every student.
By employing AI tools to aid academic success, schools can address this issue. AI can analyze data from thousands of students to find commonalities in course structure, timing, focus, and previous student success. Using this information, the AI-enabled tool can send recommendations to students based on their success in prior classes, as well as the success of previous students with similar profiles. For example, students with poor attendance records for an 8:00 AM class may be guided towards an afternoon or evening timetable. Similarly, a student who struggled through their first year English course may be steered away from courses with an essay-heavy grading scheme.
From application to graduation, there is a growing demand for AI in higher education. As a way to optimize both time and budget, AI support is a key difference maker in enabling schools of all sizes to engage a wider range of students.