“A golden age of philosophy”: An interview with Toby Walsh

Toby Walsh is one of the world’s leading researchers in artificial intelligence and has spent his life researching how machines work, how they create and how they can help us make better decisions, for example in the areas of health or business.
 
The author of the recent book 2062: The world that AI made argues that we need to stop thinking of machines as competitors and much more as allies and co-creators. He spoke to the Goethe-Institut in a wide-ranging interview on creativity, human and machine intelligence and the future choices we will have to make regarding AI.
 
You recently worked with Uncanny Valley, a team of musicians and AI experts who won the first Eurovision-style, AI Song Contest. Why do you think the winning song, co-produced by man and machine, was probably better than most of the human written compositions for the original contest?
 
I think Uncanny Valley had quite an advantage as it was really a cooperation between machine and man that made the success. It wasn't purely a computer-driven creation. Equally, it was not a purely human-driven creation, it was some synthesis between the two. I think that's where the most interesting place to play will be in the future, because we've always augmented ourselves with machines.
 
It's going to be interesting to understand how we pick up computers and use them to extend creative abilities. There's no reason to suppose that we won't be able to be better musicians and better poets, by using machines. We’ve always done other things better by allowing machines to help us. Part of the challenge is that we really don't understand our own creativity, and machines as a tool may help us understand more about this elusive idea of creativity. Toby Walsh speaking at the Goethe-Institut in Sydney | © Goethe-Institut When you look at this challenge of defining creativity, where do you see the advantages that AI has over humans, the added value it brings? And where are the uniquely human traits we have?
 
One of the features of computers is how stubbornly they follow exactly to the letter the instructions you give them. You can ask a machine to try out all the possible permutations of an idea and because it's a machine and because of its speed, it will literally go through all those possibilities. Humans may not be as thorough and systematic, and might tire and miss out some possibilities. So that's perhaps one of the advantages it has.
 
But equally, it has a number of disadvantages. Machines don't have our emotions nor our consciousness. We are starting to program machines to understand human emotions, but their emotional lives are incredibly impoverished compared to humans. And so we would have to give them much richer emotional lives. We could program them with something like a counter that says you're happy or sad or you're angry.
 
Another really important aspect, which goes to much of art as well, is that we're mortal. We live finite lives. A lot of artists are dealing with the tension that this poses to us. Machines are not alive and not conscious. Even if we do get machines to make art, they don’t have the shared human experience, which includes dying as well as seeing your loved ones leave the planet. All of these incredibly painful parts of human existence, machines will never truly experience.
 
I think it's clear that machines will never really experience some parts of the artistic process because they will never be mortal and will never fall in love the way we fall in love.
 
We tend to focus on machines getting smarter – but when we allow, for example, writing tools to do the writing for us, does our command for literacy fade, along with our critical and analytical thinking? Do you see a risk of humans getting dumber as AI becomes an integral part of our lives?
 
Yes, I think that's something we have to worry about significantly. As we hand over certain tasks to machines, we get lazy, which is a very human trait. And it's clear that this will have profound consequences. We're probably the last generation that knows how to read a map because we outsource that to machines: not looking at maps, not remembering things and just relying on devices will change us.
 
We already know that this process will physically change us: remembering maps changes the physical size of the hippocampus, the part of the brain involved in spatial reasoning. I think we have to be very careful when we outsource some of these things – we might be giving up something very important.
 
When it comes to language and our oral tradition, we used to sit down around the camp fire and pass on stories from generation to generation. That was an important cultural part of our lives. We lost that as we don't remember those stories anymore. But we got something in return – something much more valuable: we got literature, we got stories that cross time and space. We've got the works of Goethe and Shakespeare and Wordsworth, and other writers from around the world from centuries and millennia ago that we would never have had if we just stuck to the oral tradition.
 
We wouldn't be able to remember these stories and they wouldn't have spoken as far and as wide across time and geography to all of us. So we got a very good deal in return in that case. But it's not clear to me that we'll always get more back in return for having outsourced things to technology. We should be thinking carefully as we handover more and more of our cultural tasks to machines: are we getting something more valuable back in return?

Toby Walsh with UNSW robot Baxter | ©Grant Turner UNSW You argue computers can do the dirty, the dull, the difficult and the dangerous and we can sit back and enjoy the finer things of life. Do you think AI will lead to more cultural creation, more cultural consumption and more experience of culture?
 
I think it's a beautiful idea that we could end up in a place where we have more free time outside of the nine-to-five drudgery of work, to focus on the more important things of life which include the arts, our families and our societies. The history of technology to date has been that the working week has gotten shorter and we've had more time to indulge in these things.
 
So it might lead to a second Renaissance, a period of flowering of creativity. It is worth pointing out that the weekend was the invention of the Industrial Revolution. We used to work seven days a week. We used to get up when the sun came up and work until the sun went down, and then went to bed. But because of the benefits of automation and the industrial revolution, we got to demand to go to church on Sunday. And then we got to demand to take Saturday afternoon off. And then all of Saturday off and now most of us get two days off a week.
 
But that's a purely human construction. There are a number of companies in New Zealand, the UK and elsewhere that have been trialing a four-day week. They discovered two interesting facts. The first is: people are just as productive. So you stop having all your bullshit meetings and you focus on getting your work done. If people are just as productive, you can pay them for four days work where you used to pay them for five days. And secondly: people are happier, they spend more time doing the things that are important to them, whether that be reading or writing, or painting or cooking or whatever it is, that brings them pleasure in their lives, instead of wasting their lives at work.
 
And if you ask people who enter their final years of life, most people don't sit down and say: “You know what, I wish I’d spent more time in the office”. People say: “I wish I'd spent more time with my family. I wish I’d spent more time doing the important things in life and not just feeding myself, clothing myself, housing myself”. Technologies might not only help us amplify our skills, they may also give us more time. And time is of course the most valuable thing we have. The only thing that one cannot buy on this planet is time. It’s the most precious resource we have. If machines can give us time, then there's nothing more valuable we can be given by them. The year 2020 has been marked by protests in the United States and around the world | Photo Credit: Nathan Dumlao / Unsplash Your books are urgent appeals for society to make choices and take action on AI. What are some of the most important choices we need to make when it comes to AI and making society a better place?
 
I think there are two key choices. One concerns equity: we're seeing increasing inequality within our society, which is further poisoning the social discord, we already have on our hands. This can't continue, it will be too destructive to healthy, functioning societies if we do not seriously address the fact that the very rich are getting much richer, and the rest of us are being left behind. That is not a recipe for a functioning democracy in the long term. We will see people rise up. A colleague of mine was saying he's not worried about the robots taking over, he's much more worried about people rising up, well before the robots get that capability.
 
The other choice concerns our fractured political discourse. And again, AI and technology is contributing to that. We see very polarised debates within society. Increasingly, our democracies don’t seem to be functioning properly. In many places, we seem to be lacking good leadership. And that, again, is not sustainable. And we need to worry about how technology is amplifying some of those trends: the misuse of social media and machine learning used to micro-target adverts that are helping to fracture our political discourse even further. Those are the two worrying trends that we really need to address.
 
So how do you create machines that are aligned with human values such as the ones you just described?
 
That’s a very tough question. This is a really interesting time to be alive, because when we try and program machines to help us with some of these decisions, we go back to the idea that machines are frustratingly literal. We have to be very precise when we program them. That requires us to think very carefully about questions that we perhaps thought about in too fuzzy ways in the past. And that's why I think we may be entering a golden age of philosophy, in which we get to think more precisely about questions such as “What does it mean to be fair in giving out welfare or in deciding insurance rates or in locking people up, if we're handing over some of these decisions to machines?”
 
It requires us to think very carefully about these ideas and will require the assistance of philosophers, anthropologists, sociologists and political scientists as much as computer scientists.
 
As a technologist I’m only one of the people that should be trying to think about these questions. These are very deep, important questions that concern all of society. That's one of the messages in my book: all of us need to be in this conversation. It's not just the geeks in Silicon Valley building this technology, as it has been in the past. They reflect our fundamental values and are very rich and important conversations – and they are not ones with easy answers. COVID-19 could offer society the chance of a reset, says Walsh | © Colourbox Crisis and challenges can be powerful catalysts for change. Do you think the coronavirus pandemic will provide the necessary shocks to reform society for the upcoming revolution brought about by AI?
 
If we go back to the world we were at, then there's no hope for humanity. You have to realise, you can't stop the planet literally turning on its axis, and then think we go back to normal. Surely this is a wake-up call saying “What kind of society do we want it to be?”
 
I'd like to be part of a society where we value our elderly people, the people who work in the health care system, the people who clean our offices, the people who run our public transport system: all those people who have not been adequately valued in the past. So I'm hopeful that this is a moment for change, that we are thinking radically.
 
And it's worth remembering: we've done this in the past. When we began the Industrial Revolution at the turn of the 18th century, we made some radical changes. We introduced universal education, universal welfare and healthcare in most countries. We put in a lot of structural social changes to support people and to ensure that all of us benefited from the change.
 
That process equally took some shocks. The horrors of the two world wars and the Great Depression started many of those transformations, but we did make some pretty radical changes to the way we ran our societies in order to make them a better place for all of us.
 
And so I'm hopeful that, perhaps in 50 years’ time, we will look back and say, “Well, that was the moment, the shock that got us to think in longer-term ways about what sort of society it is that we want to have for our children and our grandchildren.”
 
This interview has been edited for clarity and brevity. Learn more about Toby Walsh's views on the future of creative AI here.