Are we letting a monster or saviour into the classroom?

Examining Sal Khan’s confidence in artificial intelligence.

Krish Kandiah

Krish is a social entrepreneur partnering across civil society, faith communities, government and philanthropy. He founded The Sanctuary Foundation.

A board of experts sit at a table against a conference backdrop.

Sal Khan, left, at an AI summit.

White House via Wikimedia Commons.

I've watched enough dystopian movies to know that there are lots of reasons to be nervous about the rise of the machines. Whether it’s the Terminator universe where the internet becomes sentient and creates autonomous robots to eradicate humanity, Neo battling an artificial intelligence that enslaves humans in The Matrix, or Will Smith fending off helper robots bent on taking over the planet in I, Robot, technological advances often fuel an array of nightmare scenarios. As if to make matters worse, science fiction has an uncanny knack for becoming science fact – I think of how shows like Star Trek accurately foretold mobile phones, wearable tech and virtual assistants. The line between imagined catastrophe and reality might be thinner than we might like to admit.

Perhaps we shouldn’t be surprised then, that our creative industries are sending out dire warnings about the impact of the latest breakthrough technology - Artificial Intelligence (AI). Like Mary Shelley’s Frankenstein in the middle of the industrial revolution, and Godzilla in the dawn of the nuclear era, dystopian fiction is par for the course of scientific advancement. It all stems, I believe, from our deep human response to the unknown – the fear instinct. But I have recently come across a surprising new voice of reassurance in Sal Khan’s book Brave New Words: How AI Will Revolutionize Education (and Why That’s a Good Thing).

Khan’s book comes recommended by Bill Gates - a reliably voracious reader and one of the founding fathers of the global information technology revolution. But Khan also has his own excellent credentials. From tutoring his niece online using a simple online drawing programme called Yahoo Doodle, he began creating YouTube videos and soon amassed over 450 million views. This led to his creation of the now world-renowned Khan Academy which has revolutionised online education. By 2023, it had more than 155 million registered users, with students spending billions of hours of learning on the platform.

Teachers are concerned that AI could undermine their expertise, much like satellite navigation diminished the skills of London Black Cab drivers.

You may also like

It seems to me that AI has the potential to upend the Khan Academy business model, however Khan does not take the opportunity to discredit AI or even to highlight its dangers in a bid to reinforce the advantages of his existing products. Nor does he buy into the doom and fearmongering about the impact of digital technologies on young minds, as Jonathan Haidt does in his recent bestselling book Anxious Generation. Instead, he writes a hopeful and imaginative book on AI’s potential for further transforming education for good.

Khan’s perspective comes amidst great fear in educational circles that generative AI will mean the end of education. Students can currently ask ChatGPT to generate an outline for them for an essay, suggest copy, check grammar and accuracy, offer improvements, translations, and factchecks, as well as write a conclusion, edit for wordcount, add footnote references and more. Indeed, entire books available for sale on Amazon have been allegedly written solely by AI. Teachers and lecturers are understandably concerned about the potential for plagiarism. If teachers are no longer able to discern what a student has written for themselves and what a computer has generated, the assessment process becomes meaningless.

Teachers are concerned that AI could undermine their expertise, much like satellite navigation diminished the skills of London Black Cab drivers. After years of mastering 'The Knowledge'—an arduous and demanding process requiring exceptional memory and recall—this once-essential qualification was rendered almost obsolete. New drivers now need little more than a GPS and an Uber account to compete, a shift that highlights how quickly hard-earned skills can become irrelevant in the face of technological advances. Many teachers fear a similar fate as AI continues to encroach on their domain.

While AI may not be the evil monster that will destroy us, neither is it the perfect saviour that will solve all society’s ills.

Khan offers an important alternative view. He sees the possibility that AI could, for example, help coach students on essay writing. By reading work, marking it and suggesting improvements, AI could not only save the teacher valuable time but help students take their work to an even higher level.

Khan offers a similar hopeful alternative to those who blame digital technology advances for the crisis in young person’s mental health. What if AI could help offer coping mechanisms, coaching and tailored advice that can help improve the mental health of students? His vision for the Khan academy virtual assistant ‘”Khanmigo” reminded me of BayMax from Disney’s Big Hero 6 – the large inflatable, huggable robot with a calm, compassionate and loyal personality, highly committed to every aspect of his user’s wellbeing.

Amid voices that demonise AI, Khan’s is a useful antidote, however I wonder if he has gone too far. While AI may not be the evil monster that will destroy us, neither is it the perfect saviour that will solve all society’s ills. Understatement is not Khan’s strong point. Instead, sometimes he becomes so carried away in excitement that I feel his book begins to sound like an infomercial for his own, current and future products.

I wish that Khan had taken a slightly different tack – no less inspiring about the potential of AI, but also recognising its limits. After all education is as much about transformation as it is about information. It should lead to character formation as much as skill acquisition. Emphasising these aspects of moral and perhaps even spiritual mentorship, we can see that education remains irreplaceably human.

AI has huge potential to help and to hinder us in our educative responsibilities to the next generation– and so questions remain – not if AI will change our world, but how. We need to ask not just what benefits it could bring, but who it could benefit most usefully.

Review

Books

Culture

Digital

24 September 2024

4 min read

Filterworld: algorithmic anxiety is flattening our culture

The rule of vanilla lets our unfeeling gadgets decide what’s best for us.

Simon Burton-Jones

Simon is Bishop of Tonbridge in the Diocese of Rochester. He writes regularly round social, cultural and political issues.

A podcast guest speaks in front of a mic.

What's next on the playlist?

Sebastian Pandelache on Unsplash.

Here’s another diagnosis to add to modern malaise: algorithmic anxiety. It’s described by Kyle Chayka in his excellent book Filterworld (Heligo Books, 2024) as the:

…awareness that we must constantly contend with automated technological processes beyond our understanding and control, whether in our Facebook feeds, Google Maps driving directions, or Amazon product promotions.

We don’t understand algorithms. Even if we did, we wouldn’t know how they actually work on us as every tech company keeps it a secret, lest competitors learn from them. This has led to the algorithm becoming the century’s newest bogeyman, a phantom we can reference in conversation to make us sound tech savvy and culturally knowing even while we remain in the dark.

‘Algorithmic has become a byword for anything that feels too slick, too reductive, or too optimised for attracting attention’.

Kyle Chayka

One of the oddest outcomes of the ascendency of the algorithm is the seemingly diametric effects on politics and culture. In politics it has polarised people, sorting us into opposing camps and then ensuring we hear only good things about our ‘side’ and only maddening things about the ‘opposing’ side. Instead of calmly listening to a different view, we hurl insults, as performative as Prime Minister’s Question Time and about as enlightening.

Something different is happening with culture. Here, the algorithm makes culture more homogenous; in the words of Kyle Chayka, it is ‘flattened’. The basic rule of what he calls Filterworld is that ‘the popular becomes more popular, and the obscure becomes even less visible’. It is a strange re-mix of Jesus for the digital age: ‘to all those who have, more will be given…but from those who have nothing, even what they have will be taken away.

The life of an Instagram post is said to be determined in the first five minutes. If it has engagement, it can be sure of more; if it gets none, it will sink. Visibility on social media is vital for artists of all kinds, because this is where all publicity begins. Artists try and game the system, figuring out what kind of content the algorithm will promote. In the process, their creative expression is subtly compromised. People begin to write in a style that gets attention, and what gets attention is decided by the algorithm. Those who tweet will know how the short, pared back medium starts to influence their life away from X. Musicians know that art which is safe and mainstream – the public’s crowded middle where performers like Ed Sheeran have thrived – is likely to succeed.

‘Much of culture now has the hollow, vacant feeling of having been made by algorithm’ according to the cultural commentator Dean Kissick. Chayka observes that: ‘algorithmic has become a byword for anything that feels too slick, too reductive, or too optimised for attracting attention’.

It is often at the margins that breakthroughs emerge; art that makes us see this world in a new and divine light.

There is a valid counter to this development. Previously, what we read, heard and saw as cultural consumers was determined by a small set of experts who filtered content for us. These experts were often drawn from a narrow section of society who inevitably brought their own biases to bear. While this may be true, it is hardly a triumph for the public to have an unfeeling gadget decide what’s best for them, based on what we have liked before and what seems to appeal to most people. At the ice cream vendor, this is like reaching for vanilla every time.

The truth is, in necessarily surrendering to the algorithm (for what alternative is there online?) we miss huge volumes of culture that might appeal to us. It is about as effective as deciding what sea life we like based only on what pops up to the surface of the water.

The best art is not always the most popular and there is a risk that the divine spark of invention that the creator God has placed within each of us – the unlimited potential of being made in the image of God – will not be fanned into existence as often as it could be. Chasing likes is no substitute for patient inspiration. It is often at the margins that breakthroughs emerge; art that makes us see this world in a new and divine light.

‘Behold, I am making all things new’ says the one who sits on the throne in Revelation. That algorithms are making all things similar is the reality we are learning to live with.