You have /5 articles left.
Sign up for a free account or log in.
The era of AI management of big data is upon us. Even in the slow-moving field of education, big data is making a difference. Identifying trends that never before had been uncovered and using predictive analytics to generate models of the future are commonplace in all aspects of higher ed including tracking and predicting enrollment demographics. Much more of the text generated in our field than many realize is actually written by—or inspired by—AI! Sports reports, press releases and annual reports use AI in a variety of ways. GPT-3, the third-generation Generative Pre-trained Transformer, is a neural network machine learning model trained using internet data to generate any type of text. Developed by OpenAI, it requires only a small amount of input text to generate sophisticated machine-generated text.
AI research assistants are as prevalent as Siri and Google Assistant. More sophisticated assistants such as Elicit, Leo and Wizdom offer specialized assistance in academic research. In time, digital assistants will modify the way we set expectations for research reports in our classes. Already, Walden University’s “Julian” is an AI-powered tutor that is available to assist students round the clock.
Of course, the Jill Watson teaching assistant pioneered by Ashok K. Goel of Georgia Tech has evolved into an adaptable, anthropomorphic AI tool that has the potential to take on many of the common tasks of professors such as grading, providing feedback and answering student questions. The original version took some 1,500 person hours to program. Now Goel reports he and his team can create a new “Jill” in 10 hours, creating the potential to make an AI teaching assistant available at all levels of teaching and learning.
That reduction in build time is thanks to Agent Smith, a new creation by Goel and his team. All the Agent Smith system needs to create a personalized Jill Watson is a course syllabus and a one-on-one Q&A session with the person teaching it. Named after the self-replicating character in The Matrix, the Agent Smith program clones Jill Watson, but makes her a specialist in the area of need. Teachers from any grade level or subject domain can have a deployable, AI-powered teaching assistant for their class with minimal set-up. “In a sense, it’s using AI to create AI,” Goel says, “which is what you want in the long term, because if humans keep on creating AI, it’s going to take a long time.”
In a more global context, AI has the potential to remake our class-based teaching and learning paradigm. For hundreds, even thousands, of years, formal education has been based on teaching in classes of students. The economy of going beyond private tutoring to teach a group rather than one individual is obvious. The opportunity for group exchange and engagement is also clear. However, there is ineffectiveness inherent in having to spread the teaching across a range of students with varying knowledge, skills, interest and desired outcomes. We will be better served by personalized learning opportunities.
I taught my first college course half a century ago in 1972. Many years, thousands of students and hundreds of courses later, the challenge remained the same throughout my teaching career: how to meet the needs of all of the students, not just the best prepared, not just the middle of the pack, not just the less prepared, but rather all of the students in the class. It has always struck me as ill conceived that we teach 20 or 30 students at a time. I understand the economy of doing so, but not the pedagogical sense.
It is as if we had 30 students of approximately the same age, and we gave them all size 7¾ hats and size 10½ shoes. That fits my rather melon-size head and average-size feet pretty well. However, I doubt that they would fit more than 10 percent of any 30-student class taken at random. The other 90 percent would not be comfortable in either the hat or shoes. It just wouldn’t be a good fit for most of the learners. So, why then do we teach our 30 students in the same way with the same content and teaching methods no matter how well or poorly prepared they are for the class? Can we expect that approach to make everyone come out of the class with mastery of the subject?
Sure, we do a nip and tuck in the delivery, tailoring the message just a bit for those who are lagging behind or who are accelerating ahead. Yet, these are ad hoc measures done on the fly for the most part. We suggest review sites; we give opportunities to redo assignments; we encourage advanced problems for those we believe can handle them. These are, at best, making do to correct mistaken assumptions.
An AI-powered adaptive model of teaching includes frequent probing assessments throughout the learning term, analyzed by AI to determine the depth and breadth of knowledge of the learner in the topical area and adapt the syllabus to achieve mastery of the outcomes at the end of the term. Gaps in learning can be identified in these assessments, and adjustments can be built into the flexible, personalized syllabus for the student. Areas of prior learning mastery within the subject area can also modify the syllabus for a more efficient, brief review of that topical area. Options can be offered to either shorten the class term or to fill those spaces with discussion leadership and engagement opportunities with others in the class.
There are a whole host of ways in which AI can improve learning outcomes, lessen the workload on faculty and staff, and ensure that our learners are getting the best, most relevant education possible. These outcomes do not happen without intention or proactive preparation. Who is leading the effort to apply the best AI technology in the most useful way at your university? Are you committed to continuous improvement and efficiency through new technology? How can you help ensure that your institution is on the forefront of using technology to support your students?