The role of artificial intelligence in e-learning in HE
A major asset AI brings to e-learning, either at university or within a corporate organisation, is the possibility offered to students and learners to ask questions and get answers from the technology. By asking AI all sort of questions when the learner interacts with the course content, time is saved for more in-depth classroom discussions with the instructor.
Furthermore, AI is seen as ensuring a level of privacy, as the learning process becomes, to a certain degree, a private exchange.
Adult learners and students sometimes tend not to ask questions in a group setting, fearing their peers will disapprove of what they say. In fact, evidence shows that adult learners are much more likely to ask questions and get answers from the AI because this kind of interaction is similar to a private conversation, which is not stigmatising.
Learners are much more comfortable asking the AI even trivial questions, but which they would not have done in a group setting. As a result, the classroom or group discussion can be reserved for richer interactions and matters of more substance.
As the student or learner familiarises themselves with the AI, the learning process could also define pathways through the education process. Through the types of questions already asked by the end-user, AI can help direct the learner to the next level they need to get to.
And not only where to go next in their learning journey, but also how to get there in line with the individual’s learning style and the knowledge they have already acquired. Identifying the individual’s learning style through their questions is definitely another advantage of interacting with the AI.
When it comes to contextualisation, AI enables the learner to deepen their understanding of the learning content by bringing different concepts – or any alien reality, be it geographical, historical or other – into the learner’s social context and particular reality.
AI’s role is not just about finding and guiding the student to the information they need. It’s much more about making sure the user is getting information that is understandable and has the knowledge about the context they need to understand the information requested.
Instructor’s role in an AI-based learning context
AI’s role in education is commonly seen, at least for now, as an efficient auxiliary to instructors. AI can make mistakes. And one role for the instructor is to correct those mistakes. When the AI makes a mistake, it’s often a mistake in terms of understanding of human expressions. And humans can help shape that understanding.
It might be that the AI mistook the context in which it is operating, and therefore made a mistake based on context. It might have made a mistake just because it got confused about things that people generally take for granted.
In sum, making sure that the AI interface is something that a human can work with is really important, thus reinforcing the teachers’ role. Moreover, when interacting with the end-user, the AI can uncover gaps in the curriculum, hence allowing the instructor to fill in. Overall, mentoring, leading the discussions and getting students to understand the right problems are all areas where the instructor is much needed in an AI-based learning context.
The issue with public domain knowledge
When it comes to AI’s knowledge base, it mostly relies on large public domain knowledge spaces. That knowledge is, unfortunately, often incomplete in at least two different dimensions.
The first one is when it comes to day-to-day experience. As Walter Bender, former executive director of the MIT Media Lab, frames it: “We can find an article on a car door on Wikipedia, but not so much on how to open a car door. And we are less likely to find an article on how to sit down at our desk. This unlikeliness holds true for various areas of common knowledge.”
The second dimension where AI falls short is when it comes to very specialised knowledge. For instance, an area of knowledge specific to government operations, a terminology used within the corporate knowledge of a particular organisation or a knowledge specific to a particular domain or a part of it.
Since AI’s knowledge base doesn’t primarily draw on such a specialised body of knowledge, but mostly from public domain knowledge, it would be difficult to expect AI to find answers to some particular questions. That’s another place where AI typically gets confused and makes mistakes.
AI’s epistemological stance
Artificial intelligence knows and processes the information to which it has been exposed or has been trained to process. As such, if the AI is exposed to a particular perspective, it will naturally adopt that particular perspective.
Questioning AI’s perspective or its epistemological stance is an essential step towards ensuring transparency and confronting a major ethical issue. It’s important to make sure that users see why the AI is making a given decision and to understand its decision-making process.
Transparency also applies to how the learner can learn from the AI through understanding the process it takes to come to a result. Not only is it important for the learner to have correct results when interacting with the AI, but it is also important for the learner to understand why and how it gets these and to be able to learn from the AI.
Learning on the job
When it comes to learning on the job – as opposed to through structured courses, one of the main roles AI can play is to leverage just-in-time access to information that has previously been learned, making it available when it’s needed. In this regard, AI’s role is to be a resource in the doing process because, when the information is needed, it is unstructured compared to the formal structure of other learning content or course content.
All in all, in a world where students and professionals need to learn constantly and need to adapt quickly, there is an urgent need to be able to access expertise and streamline the learning experience. AI fills this need at university, across government and in the private sector, either when it comes to formal learning or learning on the job.
Dr Dodzi Amemado is a researcher in the area of e-learning in higher education. He has recently been a visiting scholar at the Center for International Higher Education at Boston College, United States. He is a senior analyst at the Privy Council Office, the department of the Prime Minister of Canada. This article is based on research interviews with researchers at university and in the private sector. Special thanks to Walter Bender, former executive director of the MIT Media Lab, currently chief technical officer, Sorcero.