AI and the future of education are closely interlinked, and the technology is already here. The teacher in front of you may still be an actual human (or pretty close, at least) but behind the scenes, a lot has changed already. The same tech that helps Spotify recommend songs in your playlist, for example, is already in use to help students choose their next modules and courses in online learning environments.
Last week we looked at the application of AI to create a more personalized and enriched experience within the field of Education Marketing and Admissions. Beyond this, where might AI and the future of education lead us? How might it make the teaching and learning experiences more fulfilling, more effective, and more personalized? Let's look at 4 practical examples of things happening right now that we predict will become the norm very soon.
There is oh so much more potential here. Most research in this area will throw up the name David Kellerman, and this is where it gets interesting. Kellerman is really using AI in his Mechanical Engineering courses, and showing the world that AI is not just about data mining and algorithms, nor is it about far off embodied technology where machines have developed a human level of cognition.
Right now, Kellerman is making lives easier for teachers and better for students, and it's all thanks to AI. In class, he is your typical tech aware teacher, using shared digital screens and collaborative teaching and learning tools to provide a richer experience. But it's outside the class where things get seriously impressive.
In the digital platform, an AI question bot (QBot) is at work. This bot has scanned every single interaction in the platform, every second of video recordings of the classes, and even recognizes images like schematics and diagrams. When a student types in a question, the bot will try to answer it. Maybe it will respond with an answer that was already there if the question has already been asked (we know how easily things get buried in chat forums). Perhaps it will send a timestamped section of the video in response, where the question is addressed in a recorded class. If it can't answer, it will notify a (real) Teaching Assistant who is best placed to answer.
This way, the TA has the time to actually answer the question properly, instead of spending hours repeating info or directing students to sources. The quality of responses is incredible, and student satisfaction has jumped to 99% in his classes.
If you are an accomplished digital native, you can access and deploy the open-source version of the Qbot, or just listen to Professor Kellerman explaining how he uses it here. Just imagine how much this technology could improve the online learning experience right now when learners need access 24/7 to content and guidance across time zones. It is surprising that this technology hasn't yet been adopted as standard.
The tech here is less flashy than Kellerman's QBot, but for the survival of institutions in the future, a personalized approach to student retention is imperative.
Studies vary, but to give you an idea, just over 15% of students dropped out after their first year at US colleges in 2015, and in the UK that number was about 7% for the same year. The cost of this to an institution is huge, considering not only the loss of tuition fees but also the huge cost of recruiting those students in the first place.
We also have to think about the students that leave. Do they feel like they "dropped out" and that the fault was theirs? What stories do they tell themselves that may follow them through life and inform future choices? Some students just realize that a course of study is not for them, and that's actually a good thing, but losing students for other reasons is just bad news all around. When academics had figured out the behaviors that students displayed before dropping out, they were able to program that into an AI system at Ivy Tech, which then identified the students at risk of leaving and flagged them for personalized support. The dropout rate was cut immediately by 3,100 students (out of just under 17000 the year before), and the program continues to be refined and its success rate improved each passing semester. Sheila MacNeil of Glasgow Caledonian University recently said that she was far more interested in using AI to improve the quality of teaching and learning and not just focus on improving "the business of education". There is no reason that AI can't tackle both, and that's good news for everyone.
OK, perhaps the word "coaching" is a bit far fetched for AI, but that's the word used by The University of Michigan in their Ecoach program. It is used largely in STEM (Science, Technology, Engineering & Maths) classes at the moment, tracking student progress as they work their way through various tasks, problems, and projects. Ecoach learns the bits that students find easier and more challenging and prepares them for the tricky bits with extra support. It also suggests other resources and areas of interest if a student displays an interest in something particular.
Shadow Health is another example of what is known as "adaptive courseware", and one which is genuinely impressive. Using a "conversation engine", which is AI-powered, Shadow Health lets medical students practice clinical interactions with simulated characters. For example, "Doris" arrives at the hospital with pneumonia and the medical student needs to carry out an investigation and recommend a course of treatment. The AI engine is checking whether the information is covered, if the tone is professional and if the right questions are asked. It can't teach doctors to have better handwriting, sadly, but you have to admit that this is pretty special technology.
Up until now, the advance of AI into education has been quite delineated. By this, we mean that "automatable" tasks get automated and "human" tasks stay human. The passive end of knowledge acquisition can be machine-supported and learner-directed, while humans deal with conceptual stuff like critical thinking and creativity. That line is now starting to blur, especially as education as a model gets disrupted, and that's where things get interesting.
Take for example a new application called Packback, which monitors discussion boards and tracks not only engagement from students, but curiosity. This tool not only moderates entries and provides real-time feedback to students on writing effective posts, but also identifies which posts are likely to be challenging, thought-provoking and curiosity inducing, so it can put those to the top of the thread. Now Packback is presumably using a more advanced version of the tech that Grammarly use to tell you if your email sounds bossy or polite, but Hayley Sutherland at IDC predicts something that goes further. AI That can tell how you feel from your facial expression! The Centre for Neuroscience at Duke University can already do this in quite a reliable way; using machine learning to connect human microexpressions to emotional states. Can you imagine being in a lecture where the professor says something that sounds way too complex to follow? Your face registers confusion and up pops a bot on your screen: "you appear confused. We have flagged this point for follow-up discussion in the forum".
We don't have the answers. While it is crystal clear that anything which can be automated certainly will be, the question is where the limits lie. Much of the discussion of AI and the future of education centers around risks; privacy, loss of human connection, labeling students and categorizing them, etc. These are all very real and certainly valid. However, let us not lose sight of the tremendous benefits that are within reach.
More personalized learning and more learner-centered collaborative working are what the new trends in education demand and AI is perfectly placed to support it. Teachers with less stress from routine tasks and more time to facilitate deeper learning and human interaction; all of this is just around the corner. As experts in automating the marketing and admissions end of the process, we at NEO Academy are keeping a keen eye on technological advances that will support our clients to offer students better service. However, as passionate advocates for a new paradigm of education that gives learners the experience they deserve, we think AI can and should be a central part of that. Keep an eye on our blog for the latest news and ideas around the future of education and contact us to chat, discuss, or engage.