Five things you should know about AI and learning
The robots are coming – or have likely already arrived – to a school or university near you. Should you be worried?
It’s easy to understand why some feel that applying AI to learning will dehumanize education. But the algorithms and models that drive these technologies form the basis of an essentially human endeavor. AI can provide teachers and learners with tools so they can see not only what is being learned, but also how it is being learned. It also has the potential to help learners develop the knowledge and skills that employers are seeking.
So before you get worried, here’s five things you should know about AI and learning.
AI helps us open up the “black box of learning”.
AI involves computer software that has been programmed to interact with the world in ways normally associated with human intelligence. To do this, it relies on both knowledge about the world, and algorithms to intelligently process that knowledge. This knowledge about the world is represented in “models”.
There are three key models used in applying AI to learning:
- the pedagogical model, representing the knowledge and expertise of teaching,
- the domain model, representing knowledge of the subject being learned, and
- the learner model, representing (of course!) the learner
These models develop – that is, they get smarter – over time as learners interact with them.
By using AI in learning, we hope to “make computationally precise and explicit forms of educational, psychological and social knowledge which are often left implicit.” In other words, AI might be our most powerful tool to open up what is sometimes called the “black box of learning.”
When a learner uses an AI-driven tool for learning, the result is a deeper, and more fine-grained record of how learning actually did or didn’t happen. For example, AI can help us see and understand the micro-steps that learners go through in learning calculus, or the common misconceptions that arise. These understandings can be used to improve learning technologies, and they can also be used to good effect by human teachers.
AI will not replace human teachers
Although AI is designed to interact with the world through capabilities and behaviors that we would think of as essentially human (for example, by recognizing and reading handwriting), it’s important to remember that AI doesn’t have a mind of its own, and that artificial intelligence is different from human intelligence. And, as we have learned from the world of competitive chess, it’s the combination of human plus machine that provides the ultimate advantage.(cite)
So rather than a future in which AI replaces teachers, we predict that the continued introduction of AI-powered learning tools will serve as a catalyst for the transformation of the role of the teacher. Drawing on the power of both human and artificial intelligence, teacher time will be used more effectively and efficiently, and their expertise will be better deployed, leveraged, and augmented.
AI will help us better understand soft skills.
We know from our past Future of Skills research, and the work of many others, that the combinations of knowledge and skills needed for success in the future will be different from what is expected today. Although technical skills will be increasingly important, equally, if not more valuable will be three categories of “soft skills” that will be at the core of human-machine synergy:
- skills related to our ongoing ability to teach and to learn
- skills related to understanding, navigating and adapting to complex systems, and
- skills related to creativity and originality.
Going forward, AI will be able to assist educators and learners with two important tasks. First, to establish more rigorous and systematic ways of categorizing and assessing the soft skills that students acquire. (In other words, to develop hard metrics for soft skills.) And second, to understand and document the teaching and learning strategies that best help learners to develop and strengthen specific soft skills in a more structured, systematic and deliberate way.
The increasing range of data capture used by AI-powered learning tools – such as biological data, voice recognition, and eye tracking – will provide new types of evidence for currently difficult to assess skills. For example, a practice-based learning experience that incorporates elements of problem solving or collaboration might be assessed using a combination of data sources including voice recognition (to identify who is doing and saying what in a team activity) and eye tracking (to explore which learner is focusing on which learning resources at any particular moment in time).
In addition, the increasing use of AI-powered learning tools will enable the collection of mass data about the teaching and learning practices that work best, enabling us to track learner progress against different teaching approaches and, in turn, populating a dynamic catalogue of the best teaching practices suited to the development of different skills and capabilities, across a range of environments.
AI will help us bring intelligent, personal tutors to every learner.
One-to-one human tutoring has long been thought to be the most effective approach to teaching and learning (since at least Aristotle’s tutoring of Alexander the Great!). But until now, one-to-one tutoring has been an unreachable goal. Not only are there not enough human tutors; it would also never be affordable. How can we make the positive impact of one-to-one tutoring available to all learners across all subjects?
This is where Intelligent Tutoring Systems (ITS) come in. ITS use AI techniques to simulate one-to-one human tutoring, delivering learning activities best matched to a learner’s cognitive needs and providing targeted and timely feedback, all without an individual teacher needing to be present. Some ITS put the learner in control of their own learning in order to help them develop self-regulation skills; others use pedagogical strategies to scaffold learning so that the learner is appropriately challenged and supported.
One new example of an ITS is Aida, Pearson’s AI-powered mobile calculus tutor. In Aida, AI is applied to multiple tasks to help personalize the tutoring to the student’s learning path and capability. Specifically, AI is used to recognize and analyze the student’s handwriting and problem being solved; analyze each line of a written solution; provide step by step feedback on what is correct or incorrect; give intervention hints about how to solve a problem, including relevant explainers or examples; and recommend other related concepts to practice or learn in order to strengthen understanding.
AI-driven learning companions will support our lifelong learning journeys.
The early versions of Learning Companion Systems, developed in the 1980s, were collaborative computer-based learning companions. Companions used collaboration and competition to stimulate student learning. They could also act as a student for the human learner to tutor, and in so doing helped the student learn.
The next generation of learning companions are poised to become an essential part of lifelong learning. There are no technical barriers to the development of companions that can accompany and support individual learners throughout their studies – starting with school and extending through higher education, multiple careers, and retirement. These lifelong learning companions could be based in the cloud, accessible via a multiplicity of devices, and operated offline as needed.
Rather than teaching all subject areas or skills, a learning companion might access specialist AI systems or humans as needed. In addition, the companion could focus on helping learners to become better at learning, for example through prompts focused on developing a growth mindset. And because this type of system can help all learners to access learning resources that are optimal for their needs, it will be suitable for struggling learners as well as those who are high achieving.
To make lifelong learning effective and successful, we know that learners will need better navigational tools and services to map their learning path. A lifelong learning companion could also fulfill this role. Think of this as a Waze for learning. The companion could help learners identify their strengths and weaknesses, and suggest methods and resources to build skills more quickly. And, it could be combined with sophisticated labor market analytics yielding more granular insights into how jobs, skill requirements, and career opportunities are evolving. This information would help learners to identify career opportunities, acquire new knowledge and build requisite skills.
About the author:
While at Pearson, Laurie Forcier directed Pearson’s program of Global Thought Leadership, where she translated world-class education research into forward-looking pieces that are practical and accessible, creative and inspirational. Laurie has nearly 20 years of experience in the education sector, covering research, evaluation, policy, and administration. She previously worked at the Harvard Graduate School of Education and at the Urban Institute, a Washington, DC-based think tank. She is the co-editor of Improving the Odds for America’s Children: Future Directions in Policy and Practice (Harvard Education Press, 2014), and co-author of Intelligence Unleashed: An Argument for AI in Education (Pearson, 2016).
This blog draws from two recent Pearson thought leadership papers Intelligence Unleashed: An Argument for AI in Education (Luckin, et al., 2016, Pearson) and Opportunity for Higher Education in the Era of the Talent Economy (2019, Pearson).