Artificial intelligence sounds like the stuff of science fiction, but in actual fact we’re downloading apps every day for our phones that use it. For instance, Google Photos uses AI to classify photos into common categories like “cats”, “trees” and, yes – even “robots”. As a result, we probably already use AI in education far more than we realise. And when it comes to robots, whilst we might think of them as something futuristic, keep in mind that Amazon has stealthily deployed around 30,000 robots in its distribution centres, so in a sense they are already here to stay.
AI and machine learning could really come into their own in education because they offer us the scope to adapt the curriculum, based on real-time feedback from teachers and leaners. Technology enhanced learning is another key topic on the programme at this year’s Digifest. As innovative as it can be though, the tech won’t work unless it’s accompanied by pedagogy, something that very much runs through Jisc’s approach to supporting colleges and universities in their engagement with technology.
We’ve seen some quite expensive education technology in the past, often with a steep learning curve. What’s great about much of the ed tech we’re seeing in the marketplace right now is that it’s very affordable for schools and home educators. For instance, Google’s Cardboard virtual reality (VR) headsets retail for around £5-£10, transforming bog standard phones into magical tools that give learners the ability to take an underwater safari, visit the pyramids of Giza or see what the world looks like from space. Yes, you can use your phone in this lesson!
For those who have concerns that our society, and our education systems, are becoming too reliant on tech, it’s worth pointing out the learning gains that can be made for students engaged with this technology – and teacher engagement is key here, too. In a recent Jisc survey, more than 70 per cent of students said their understanding of technology was greatly enhanced by their teachers and lecturers leading the way. But it’s a big ask to expect teachers to bone up on VR, AI, 3D printing and robotics. As we’ve seen, the new computing curriculum is proving quite challenging to implement.
Nurturing the innovators of the future
Perhaps we can get some ideas about how this could work by looking at how other countries are addressing digital skills and digital capability. For example, in Italy, the ministry of education is creating a makerspace in every school across the country. Perhaps this is an initiative that could be adopted more widely in the UK as part of implementing the government’s digital strategy – giving young learners hands-on experience of the technologies that industry so desperately needs. If we can foster and nurture the maker, hacker and tinkerer, then this will go a long way towards addressing today’s skills crisis – whilst perhaps also creating the right conditions for the entrepreneurs and innovators of the future to flourish.
Going back to AI, Google’s Deep Dream project is a fascinating application of the technology. Deep Dream uses artificial neural networks, a style of computing inspired by the brain and nervous systems. Neural networks are widely used to recognise shapes in pictures, trained from massive public and private datasets. Deep Dream turns this on its head, and uses the neural network to mutate an existing picture into something that’s, let’s say, a bit more cat-like. Lo and behold, cats start to appear…
Lately we’ve also seen examples of “adversarial” neural networks, where one network generates images, and another scores them based on its knowledge of the subject area. Let’s picture that the first neural network tries to make a cat. It doesn’t know what a cat looks like, but the second neural network will tell it how well it did. After a few hundred thousand “cats”, it starts to do pretty well at creating its own cat pictures.
What does all this have to do with education? Well, if we can generate cat pictures, then perhaps we can generate all kinds of other things. Already we are seeing AIs writing music, books and all sorts of other creative activities. Perhaps the key lesson here for education is that the learner of tomorrow might be working from a unique personalised curriculum, generated to their very specific needs. How do we go about assessing learning gain in this situation? This is going to be a very real challenge, and something that we are already starting to see with some of the AI powered ed-tech apps I saw at this year’s Bett show.
Perhaps the most exciting thing about today’s AI is that much of the underlying technology has been made freely available for learners, educators and entrepreneurs to build on. For example, Google’s open source TensorFlow system powers the AI in tools like Google Photos, and has quickly established itself as the platform of choice for AI developers worldwide. For young questioning minds, this can only be a good thing.
Martin Hamilton is Jisc's resident futurist. The Jisc Digifest runs from 14-15 March at the ICC in Birmingham. TES is media partner for the event.