The future of education: a new way of learning


Technology is about to transform the world of learning as we know it, writes David Rogers

The driving force behind educational transformation is information and communication technology: a combination of computers, software, wireless connec- tivity and abundant personalised online resources (effectively, pupils will be able to have a library each). This is qualitatively different to the whiteboards and IT suites on which schools have devoted much of their capital budgets in the past 10 years: they may have done away with the overhead projector, but they’ve not otherwise altered the structure of class teaching.

Class struggles

Over the past two hundred years or so, education for most children in most societies has been a collective experience. The reason is simply that whereas teachers are scarce and schools are expensive, children are plentiful and, so to speak, free. Starting from this given, an extensive literature has accumulated over how best to teach a class. If that sounds like a simple problem, it shouldn’t: the way the question is answered has profound social and economic consequences, and has become a bitter contest between competing political philosophies, stakeholders and fashions.

For example, a study by the Organisation for Economic Co-operation and Development carried out last year found that although the average teacher in England and Wales works a 48-hour week, they spend only 20 in the classroom. And of the time they do spend with students, little is spent on one-on-one instruction aimed at tackling a particular pupil’s particular problems. A teacher’s time is, therefore, a precious resource.

This, in turn, is connected to an issue that has dogged education: the extent to which children should be placed in groups depending on their abilities. In subjects such as mathematics and languages, some pupils are able to progress rapidly whereas others require continual support. In a class of 30, a teacher has only two minutes per pupil per hour (and only 20 hours a week), so where should their time be allocated? That’s up to individual teachers based on their knowledge of the class, but there’s no right answer because whichever choice is made, some pupils will lose out.

The obvious response to this issue is to “set” or “stream” pupils, but this is also controversial because to put a child in a low set is a self-fulfilling prophecy. In the words of a 2004 Department for Education and Skills report on mathematics courses, “many students feel themselves to have been classed as ‘failures’ by their teachers before they even start the course”. And hovering somewhere above the debate are large, nebulous issues of social mobility, equal opportunities, future industrial performance, and the preservation of cultural identities.

The books that read you

What the technological revolution promises in education is a way of replacing the approximate, averaged-out, time-poor learning of the conventional group with programmes that accurately address each student’s individual needs, and at the same time relocate teachers from the front of the classroom to the elbow of the pupil.

The initial focus of “AI and Ed” regimes has been the textbook, because it is easy to see that a smart version could take over some of the straightforward tasks of a teacher, such as making sure pupils memorise verbs and formulae and practise the manipulation of equations in maths and science.

Ben du Boulay is the emeritus professor of artificial intelligence at the University of Sussex, and president of the International Society for Artificial Intelligence in Education. He has been involved in attempts to extend the principle of the smart textbook into a learning environment in which tutoring systems try to imitate what teachers do. The aim, he says, is to build a model of the pupil’s understanding, a model of the subject and a model of the appropriate pedagogic response.

“It has to build a sense of what you can and can’t do that changes over time. So each time you successfully solve a problem it has to update its understanding of your understanding of the skill components required to solve that problem. Then it has to make a judgment as to what it, as the tutor, should do – so if you appear to make a wrong step it has to decide if it should just say ‘wrong’, or keep quiet, or give a light or a heavy hint.”

The bandwidth problem

Obviously this approach works best in formal systems such as algebra, although the ability of AI to handle translations in natural language has greatly improved. However, the real issue with making tutoring systems more effective is increasing their “bandwidth” – that is allowing them to gather more information about the learner than just their response to a problem. Du Boulay has specialised in developing systems that can understand the student’s mood and emotional responses, and then adapt its behaviour to motivate them, in much the way that a skilful teacher might.

He says: “There are plenty of people who now build systems that point a camera at a student’s face and try to judge whether they are smiling or frowning. Some instruments study the chair that the learner is sitting on to see if they are wriggling or leaning forward or leaning back, or they attach galvanometers to their hands to measure skin resistance, or they look at heart rate or they apply electrodes to the scalp to pick up brain signal.

“As yet there are some small successes, but there are lots of problems. Firstly, because it’s intrusive; secondly, there is the issue of how you calibrate the signals. These days you can reliably detect if someone is frowning, but is it a sign of concen- tration or are they slightly upset? Thirdly, if you can reliably detect the state they’re in emotionally and motivationally then what do you do?”

One response to these issues is to have the system engage with the pupil, for example by asking them if they are bored because the material is too difficult, or because it’s too easy. This process is made more effective by having the tutor appear to be a person. Many tutoring systems now have animated speaking heads, and adopt appropriate facial expressions in response to what a student has done. Du Boulay says that a lively area of AI and Ed research is into what face the system should have: male, female or that of another child – or a different face depending on the subject or the system’s understanding of the pupil’s mood and emotional state.

Some of these personalised systems have already made inroads into the educational system, particularly in the US. Here, products such as Cognitive Tutors have been used on more than a million schoolchildren. So far, most studies into the effectiveness of intelligent tutoring systems have shown small but significant improvements over conventional classroom teaching, which is a reasonable result given their relative youth.


As well as changes to the teaching material, the whole learning environment is likely to be reorganised according to new principles. The term of art for this process is the “flipped classroom”, meaning the part at the beginning of the lesson in which the teacher instructs the class is moved to the slot occupied by the previous day’s homework. This means that when the class assembles for the lesson it is ready to get down to analysing the information in the presentation, and the teacher functions as an adviser rather than a conventional instructor.

What will not change, however, is the classroom itself. There are a few in the AI and Ed field who argue that classrooms will turn into libraries and teachers will disappear, but the overwhelming view is that education will remain a collective experience. Du Boulay says: “Part of what happens in school is social, and classes are important because a certain amount of learning is vicarious – you didn’t ask the question but your friend did and you heard the answer. So, I think if we came back in 50 years schools would be still recognisable, but the use of computers of all kinds would be much more embedded and everyone would be more comfortable with blended education that moves backwards and forwards between intelligent tutoring and human tutoring.”

The possibility of, as du Boulay says, “cutting the knot” of mixed ability vs streaming is the great promise of the new classroom. That offers the hope of giving every child every chance of making the most of their school days.

Out of the class and into the woods

Some schools are taking the opposite tack to the increasingly high-tech environment of the classroom of tomorrow. The Forest Schools movement became popular in Scandinavia in the 1950s and has been taken up elsewhere as a form of technology-free experiential learning. These friluftsliv schools, as they were originally called, serve mostly younger children – four- to seven-year-olds, but there are some programmes that cater to older students.

The philosophy is to exploit the natural world as a teaching aid for biology, botany and physical geography, but also to learn woodcraft skills and to begin to create a relationship between children and the natural world. The movement is spreading through Britain, and is already popular in the US.

More from David Rogers:


Photo Credit: Redd Angelo

Previous articleHow to create a healthier planet: Be willing to be wrong
Next articleA call for a new compassionate world lens
David is a freelance journalist based in London specialising in business and politics. For many years, he worked as the chief sub-editor of a magazine in London, before going freelance as a feature writer. A passionate history buff, he is working on his first novel set during the 1870 siege of Paris.