Machines of many parts

The great privilege of being a journalist for Times Higher Education is having the reason and opportunity to interview scholars who are, in different ways, shaping our collective future.

One such researcher is Geoff Hinton, a pioneer of the potentially era-defining field of machine learning, who is renowned in particular for his work on artificial neural networks.

Inspired by the human brain, neural nets underpin many of the best-performing artificial intelligence applications we live with today, such as the speech-recognition software in your smartphone.

In addition to driving industrial innovation, however, Hinton’s research also raises fundamental questions about what it means to be human, something I asked him about ahead of the THE World Academic Summit, to be held on 1-3 September with the University of Toronto (UoT), where he is based.

“Either you’re a materialist or you’re not,” he told me. “If you believe that we’re machines – we’re these wonderful machines, and we’re influenced by the fact that we live in societies, and we learn from other people and have a culture as well as these huge neural nets – then there’s no reason why there’s something that you can’t mimic in hardware or in simulated software.

“The alternative is to believe that there’s some kind of spooky stuff, probably related to consciousness, that is not mechanical and not something you could ever put in a computer, and that just seems silly to me.”

Whether or not you agree, it is notable that this view does not discount the fundamental importance of people and culture and society, the stuff not of hard sciences but of the arts, humanities and social sciences. And yet, the popular view is that such disciplines are in decline – of less relevance to the world today, when science and technology rule.

In our cover story, we interrogate this, looking at the longitudinal data across the world and finding a picture that, while mixed, does show a widespread shift towards STEM since the financial crash.

While this is not as dramatic as some might fear, there is no shortage of scholars sounding the alarm, from linguists (who, the data suggest, are more entitled than some others to feel existential angst) to leaders such as Nick Dirks, president of the New York Academy of Sciences.

In a recent essay for THE, Dirks warned that science scepticism had risen alongside scientific advance (note Covid and climate change), arguing that coordination across the “two cultures” was more important than ever. A stumbling block is that arguments for plurality are often seen as special pleading when, as Dirks puts it, “there is a widespread view that the humanities are both largely irrelevant to contemporary life and ill suited to preparing students for careers”.

But even if one takes a utilitarian view, a focus on a small number of STEM fields at the expense of all else is a mistake.

I asked Hinton whether the multitude of disciplines commonly found in a university environment was a uniquely important attribute, from the perspective of someone working in a field such as artificial intelligence.

Not necessarily, he said – not because this diversity of expertise and perspective was not important to research but because it had to a certain extent been replicated by the most effective industrial laboratories, from the legendary Bell Labs to today’s equivalents run by Facebook or Google.

For Hinton, though, an even more pressing question is the importance of non-STEM disciplines as foundations for a decent society.

“My view is technology allows you to develop lots of goodies, but how those goodies get distributed and used depends on things that aren’t technology – social decisions,” he said.

“I recently saw a talk on something called self-healing minefields, and the idea was that the poor old minefield, when someone stepped on it, would get a gap in it and had to heal itself.

“This just seemed completely obscene to me: to talk about healing from the point of view of the mine. And that’s what you get if you don’t have enough humanities, I think.

“So whenever I give money to UoT, I give it to the humanities because I think they should get more funding. And I am very happy when universities use the big grants for the sciences to help subsidise the humanities, I think that’s a good thing to do.”

john.gill@timeshighereducation.com

Geoff Hinton will be in conversation with John Gill at the THE World Academic Summit on 1 September.