AI won’t just be in the classroom, it will be everywhere and it’s impact on education will be profound, writes Christopher Jackson
Buckinghamshire isn’t where you’d necessarily expect to find the next thing: with its Capability Brown parks and scattered manor houses, it can feel more like an encounter with the past. But the presence of Bletchley Park 20 minutes from the University of Buckingham is a reminder that the latest thing doesn’t always originate in London.
Nor is change always brought about by a radical. Sir Anthony Seldon – the vice-chancellor of Buckingham and former headmaster – has written books about five prime ministers. Knighted, fêted, he might seem the epitome of the establishment. Founded by Margaret Thatcher in 1983, Seldon’s university is at the forefront of a wave of innovation in artificial intelligence. In October it was announced that the UK’s first Institute for Ethical AI in Education (IEAIED) will open here. Seldon says the place has always thought creatively. ‘We have 39 weeks’ study a year,’ he says. ‘You get your degree a year earlier, you spend less on accommodation, there are fewer student expenses, it’s more economic, and it’s brisker.’
This creative approach has been successful – the university is expanding – and Seldon now wishes to reproduce this focus on the ethical dimension with regard to AI. Also in 2018, Seldon produced, along with one of his students, Oladimeji Abidoye, a landmark work on the topic entitled The Fourth Education Revolution.
‘AI is the Cinderella subject, for the government and for educationalists in general,’ the authors write. ‘They are still locked in the 20th-century mindset without even realising that they are. What needs changing is the mindset.’ Seldon and Abidoye describe a potential AI-enhanced world of pupils liberated from exams, and teachers freed from marking: an education tailored to individuals. It’s a refreshing contribution to a hitherto simplistic debate. Tom Stoppard wrote: ‘A scientific approach to the examination of phenomena is a defence against the pure emotion of fear.’ The AI debate tends the other way. We fear a dehumanising science – a world too machine-oriented, and less interesting. But for Seldon, technology isn’t the worry: what’s concerning is the misery of a boring education – all of which AI can help remove.
‘The factory model has squeezed out the quirky teacher,’ he says, before adding: ‘I strongly dislike rote learning as a basis for education. Too much of what passes for education is the learning of facts, and their regurgitation.’ We have become a nation of Gradgrinds. In Seldon’s view, the UK is lagging badly behind. A few weeks later, I am at the Carlton Club Dinner at the Dorchester, noting the dynamic between Seldon and education secretary Damian Hinds. Nothing could indicate Seldon’s importance more than the bowed respect with which the minister listens. I am reminded of a line in Seldon’s book: ‘The really frightening thing is that the Government is not stepping up to the mark.’ Hinds looks like he wants to do so. Also at the dinner is Tim Levy, the CEO of A Singular Life – a company founded in January 2018, which offers a unique AI experience by curating what he calls ‘non-material legacies’: the actual memories, impressions and wisdom our lives contain.
‘We are used to private client practitioners advising us on our material estate,’ he explains. ‘At A Singular Life, clients are assigned a curator, who looks at our non-material estate, and then uses AI to bring that out.’ For example, if you particularly wanted to communicate to your progeny an experience in nature – a waterfall, for instance – that could be preserved and communicated in AI. Levy is another advocate of AI’s potential to transform education. ‘If you accept the view that humanity is defined by its extraordinary capacity to create and transfer complex culture from one generation to the next – in other words, to sustain and evolve principally through education – then AI is going to be monumentally significant,’ he says. ‘Learning by rote, lecture and reading will seem archaic. Instead, learning through very vivid, immersive and AI-enabled virtual environments will become a norm.’ Levy points to the breakthrough this might trigger in history teaching. ‘In the near term we expect AI to fully replicate the thoughts, actions, mannerisms, speech and behaviour of a human being. For example, it would be possible for Vladimir Putin or Donald Trump to, non-biologically, “live for ever”.’ Why learn history from a teacher when you can learn from its protagonists?
To talk to Levy is to feel the future encroach: it’s exciting, but one can also feel bewildered by the sheer scale of change AI might bring. It’s therefore with a degree of reassurance that I arrive at Portcullis House to meet Nicky Morgan, the former education secretary, who left office in 2016. Did AI come across her desk during her tenure? ‘Yes,’ she says, in the brisk way certain politicians have of reminding you that their time is valuable. ‘Obviously, people wanted to talk to me about how AI could revolutionise the classroom. There was a big trade show every year, and it was all very exciting.’
Then she pivots: ‘At the end of the day, we’re all social beings – and there’s only so much AI can replace. Obviously, technology has its place. But teacher recruitment adverts will try and capture that moment when the light bulb goes off [in a pupil’s mind]. I wonder whether the robot hologram will have the same effect.’ Morgan does see AI delivering benefits, though, and knows her successor is focused on it: ‘Damian is looking at teacher workload,’ she says.
A Department for Education spokesman tells me: ‘Technology allows students to do a great deal from within their classroom – explore the rainforest, steer virtual ships or programme robots. Using technology, teachers can access training, share resources with colleagues and update parents on a pupil’s progress without being taken away from their main focus – teaching.’ As well as reminding me of a passage from The Fourth Education Revolution, this strikes me as more interesting than the average statement from a government institution. It is a reminder that the debate around AI is sometimes too philosophical: this one will be driven by hard reality.
One such, of course, is job market reality. Morgan sits on the board of a company which is one of the success stories of the past year: Finito, a firm committed to forging bridges between education and the workplace. Its CEO Ronel Lehmann has put together a prestigious board which, in addition to Morgan, includes Addison Lee founder John Griffin, Dame Mary Richardson, and Seldon himself.
One service Finito provides is to ensure candidates are better prepared for AI-based interviews than rivals. ‘The graduate who comes to us wanting to work as an investment banker or financial services professional is now not going to be met by a human being: they’re going to be met by a computer,’ Lehmann explains. ‘HR has set the algorithms and students have to conform to what they’re asked: it’s a cost-saving mechanism.’
Lehmann admits the difficulty of battling that: ‘You’re not going to get the questions; you can’t get a sample paper. The algorithms are set.’ But Finito still gets astonishing results through its AI training. ‘If you have 150,000 going for a job, the Finito candidate is going to stand out,’ says the CEO. ‘It’s what we do.’ One student who benefited from Lehmann’s training is Joshua Pauk, who describes the harrowing nature of an AI interview, having secured his desired job thanks to the firm’s guidance: ‘The technology behind the screen is analysing your every move, from your facial expressions and breathing pace to your clothing.’
All of which is beginning to make the transition to AI sound altogether less desirable. My conversation with Lehmann illustrates how resources-strapped HR departments make an AI reckoning inevitable. Another driver is competition arising from the global economy – from India and, of course, China. Happily, Tingting Zhong, an operations manager at the UK branch of Wailian Group, a Chinese education consultancy, sees Britain as a natural partner in AI, more so even than the US. She particularly admires our strength in university research, vibrant start-up culture, and, contraire Seldon, track record of government support.
‘China is welcoming collaborations from UK companies or individual talents for researching new AI goods and services,’ she says. ‘Wailian Group is organising a UK-China AI conference in London in 2019 which will bring industry leaders from both countries to exchange ideas and information, and discuss potential further collaborations.’ Zhong has a Seldon-esque view of the role AI should play.
‘I strongly believe children should be taught with AI, but not by AI,’ she says. ‘The practical application of AI ethical principles is still very vague and almost non-existent. We still need to find solutions to implement those principles.’ This ethical terrain is already being fought over by thinkers from both left and right.
AI is part of an education debate which can sometimes feel unhelpfully binary. The journalist and former director of the New Schools Network, Toby Young – who is compiling a report on AI for the Centre for Policy Studies – flags a third pressure which means we won’t have a choice on whether to use AI. ‘There aren’t enough teachers,’ he says. ‘Progressive politicians discuss this as if it were an isolated problem entirely due to reform programmes of successive Conservative governments.
Actually, it’s a global phenomenon.’ Young explains how AI opens up on to a debate about teaching techniques. ‘There are various objections [to AI teaching],’ he says. ‘The standard objection is that the learning process depends upon a relationship between two human beings, and that children learn because they want to please their teachers, who serve in loco parentis.’ But Young notes that children respond well to computer games, and adults to language apps: why should AI learning be different? ‘Another objection,’ he continues, ‘is that what teachers do is too complex to be done by machines.’
He admits: ‘Intelligent machines are not sophisticated enough to replicate those sorts of complicated aspects of teaching – but they soon will be.’ For Young – as for many on the right – the most effective method of teaching is direct instruction. ‘We would think of this as traditional teaching, where a teacher imparts knowledge to students,’ he explains. ‘It’s often derided by progressive educationalists as “chalk and talk”.’ He sees this derision as a left-wing fallacy.
Instead, AI can emulate this method, saving teachers from straightforward lessons and freeing them up for difficult classes. If Young is right, one wonders whether AI might, by its very effectiveness, detoxify the contention surrounding education. Where will the breakthrough come? ‘It may be that it won’t happen in the developing world as the cost of labour is so low,’ Young suggests. ‘And lots of schools in Silicon Valley use intelligent machines to teach children – but those people are rich and will do well willy-nilly. It won’t be a true test.
Proof of the concept will come probably at a US charter school or academy school in England introducing intelligent machines using direct instruction to supplement what teachers are doing.’ Whatever direction we take, it’s worth noting that the argument is dominated by people with admirable passion. Both Seldon and Morgan are driven by memories of teachers who helped them on their way.
Seldon’s was an English teacher called Jonathan Smith (‘He began by saying, “I don’t know any more about Yeats than you do”’). Morgan’s was a certain Rosemary Thynne, who inspired her to be head girl at Surbiton High. ‘I’ve got the best job in government,’ says Damian Hinds. His predecessors – Margaret Thatcher among them – presided over dusty classrooms filled with the squeak of chalk on blackboard.
Hinds’ successors will need to prepare children for a world where AI isn’t just in the classroom, it’s all around them.
Christoper Jackson is deputy editor of Spear’s