AI in Education: Utopia or Brave New World?
I was recently asked to give a lecture at Lancaster University Management School. It was great fun and a pleasure to speak with, and indeed learn from, a fascinating and engaged cohort of students and their lecturers.
One of the things we discussed was how much technology has changed over an incredibly short space of time. In a quarter of a century – in other words, in just over the lifetime of most of the students in the lecture – we’ve gone from fixed dial-up with speeds of 28k (remember when a 56k modem felt incredibly fast!?) to 5G wireless broadband, through which you can stream broadcast quality TV box sets without pause.
We’ve gone from data storage printed on paper and stored in actual physical lockers, to terabytes (which to all extents and purposes was impossible 25 years ago) of data stored virtually in the Cloud.
We’ve gone from transferring computer data, from one machine to another, on (physical) floppy disks, to seamless automated data transfers and ETL processes in the Cloud.
Now, just a few weeks after my lecture, with UK universities having suspended all face-to-face teaching, and schools closed to most children for the foreseeable future; it’s going to be a while before I, or anyone else, can join a room full of students like that again.
And just like that, in a position where we are reliant on technology; for business continuity, for a social life, and for education.
Education is already one of the biggest areas of impact in terms of technological change. When I was at school 25 years ago, Information Technology was an optional extra lesson; now it’s compulsory on the curriculum. That’s an incredible change over a remarkably short period.
There are many technologies immediately available to us on a day-to-day basis. Already we’re seeing dedicated teachers live-streaming classes, with students able to watch and participate from home.
But what longer-term changes might we be looking at, and what are some of the implications?
Technology change in education
Data technology is continuing to advance and evolve, and now we have things like Machine Learning and Artificial Intelligence; even the language is interesting, making it sound like we are more interested in teaching computers than we are in children and students.
So, given we have discussed previously the potential of AI in the workplace and how AI might be able to help companies, what about AI in education?
Some aspects of the way children are taught today in schools are still very consistent with how subjects have been taught for 30 or even 50 years. While reading schemes may change and today’s grandparents were certainly not taught to read using phonics, the basic teaching format of around 30 children in a classroom with one teacher is still the same.
And while many secondary students now have an iPad instead of textbooks, they still submit homework to be marked by a teacher and sit an exam hand-written in pen. However, it’s possible that some of this could start to change with the inclusion of AI in education.
Opportunities for AI in education
The Institute for Ethical AI in Education published a report a few weeks ago, which highlights some of the opportunities and some of the potential risks.
One particular area where much more is known today is in learning styles or preferences. Some are visual learners, and some learn by doing. For some, learning by rote (what you might call the old-fashioned way) works, and for others it doesn’t.
Much more is known about learning styles, and often children are assessed at a fairly young age as to what their own preference is, although I suspect that knowledge is often not handed from one teacher to the next. Feeding that into automated algorithms could optimise learning at an individual level.
Perhaps more straight-forwardly, AI could enable automated setting and marking of work, again with the potential with more detailed individualised feedback for each pupil. Whilst this could definitely benefit students, arguably the greater benefit would be in reducing the time spent by teachers on a somewhat repetitive and time-consuming task. This would free teachers to spend more time on lesson planning and in-classroom interaction with their pupils.
Professor Rose Luckin, one of the report authors, says, “AI is a powerful tool that can open up the ‘black box’ of learning, providing a deep, fine-grained analysis of what pupils are doing as they learn, meaning their learning can be ‘unpacked’ as it happens.”
The potential for AI to transform Ofsted inspections
AI could also revolutionise Ofsted School Inspections. Ofsted inspections are the current government gold standard, used to ensure all schools are reaching certain minimum (or, indeed, maximum) standards and to help schools identify areas where they need to progress.
However, every teacher knows that schools spend frantic hours/days preparing for Ofsted inspections and the picture the Ofsted Inspectors see is not always a true representation of the reality of the school’s daily life. At the same time, it increases stress levels for teachers for the run-up to and period of the inspectors’ visits, so can’t help with mental health and welfare!
The potential for an AI system to assess schools through a more continuous monitoring of data, based on the school’s normal activities, rather than by a specific ‘Ofsted focussed report’ could transform this. If done correctly, this could reduce the stress on teachers and schools of preparing for an Ofsted inspection, whilst still ensuring that all schools meet a minimum standard of education provision based on a more accurate data set.
The ethical questions of AI in education
Clearly there are fascinating benefits for AI in education, but, as with all aspects of AI at the moment, there are significant ethical issues still to be addressed and answered. The data being used would be highly personal and sensitive, and also about children, so there are extremely important privacy challenges.
But beyond that, there could be issues such as teacher impersonation and who has access to these systems; if it’s possible to understand what an individual’s school data looks like, could that be used to project forward to performance in later life?
Equally, the potential challenge of ‘dumbing down’ an entire generation of kids through slavishly following what a computer believes is the best answer needs to be addressed. At the other end of the spectrum, protection from gamifying an AI marking algorithm to produce the best results would need to be built in.
Could AI ever replace teachers in the classroom?
To be clear, there will also always be a place for a real live teacher in the classroom – and we must all hope that we are able to return, safely, to that norm soon. No one is suggesting that AI in education could or should replace live teachers – though, at time of writing, remote learning technologies are having a major moment.
Just as in other areas of AI development, the principle that AI should underpin and support humanity and human oversight, rather than being created for its own sake, must be a cornerstone. And many of these issues are still to be sorted out over the next few years.
However, the best indication that they will be is the amount of progress made through tech in the education sector, particularly over the first two decades of this century. Just imagine how education will advance in the next two!
AI in education is simply the next step in the ongoing progression that has led from chalk blackboards, inkwells and acetate overheads to Smart interactive whiteboards, learning iPads that can also connect wirelessly to 3G pixel wall projectors and online resources and gamification of learning.