Bill Thompson: A Talk from Another Century

Talk about the misuse of AI and the dangers it could bring if not used properly, focussing attention on the wrong things. Keeping humans at the center of the technology, not to replace. 

“We are in 2107 years in the future, we are looking back in time to warn 2024 of the misuse of AI. 

We have 2 billion less of the population, water levels rose and took land, harvest and some technology away from us and not everyone survived or knew how to do everything you do in the present. 

We should have worked together more, we do now. We could have focused on the right things to use AI for rather than to help the rich stay rich and live longer. We appreciate what value people can offer rather than making them redundant.” – Mike 🙂

Slides

Find out more about BoS

Get details about our next conference, subscribe to our newsletter, and watch more of the great BoS Talks you hear so much about.

Transcript

Bill Thompson 

Hello everyone. Yeah, I’m not what was anticipated, but I am what you get. So see how it goes. Feel free to leave halfway through, or indeed, a quarter way through, if you’ve had enough. Mark, WhatsApp messaged me a couple of days ago and said, Don’t suppose you’re free on Tuesday, and I am free on Tuesday for Mark, because we all know he’s very persuasive. When do the children come home? You said you’d release them by now.

So I’m going to entertain you until about 10 o’clock, talking about. Everyone’s talking about it, talking about AI, but from a particular perspective, and I’ll share that with you. You can find me on all the usual social networking type things. So this is the short version I’ve been projected back in time from the year 2107 Hello everybody. Coffee still rubbish to talk to you about how we got where we are in 2107 and maybe some things you might think about as we start to deploy AI technologies.

The giveaway is we haven’t sorted out all the problems by then. Let’s just leave it at that. It’s not all good, but you’ll see we’re going to start sort of with a general meditation around the way the world is changing. I got this from the Tempest, obviously.

A full AI enabled the business lies of its code, our models made those are prompts that were its teams. Nothing of it that doth fade, but doth suffer an AI change into something rich and strange. Chained prompts ring its nail. Ding dong. Hark the model speaks, what will it tell? And that’s sort of the framing that we are in the midst of this significant change, and some people are skeptical about how far it will go, and some people are skeptical about the achievability of some of the goals. But as I look around the world, I see the impact already of even what we’ve got today, imperfect though it is, and I can’t help thinking that some modest improvements in things like training speeds for large language models or the capabilities of the hardware on our phones will put this technology in everything and everywhere, and we will have to learn to live with it, even if it’s not perfect in the way we’ve learned to live with the internet, spreadsheets, the web, computers, telephones, all those other transformational things. Back down to the cataclysmic invention of writing 5000 years ago, which really did change everything.

I come to this as someone with a mixed background, as in, I studied philosophy and then psychology. In my second year of doing psychology, my supervisor for my undergraduate dissertation on computer vision systems was this PhD student called Geoff Hinton. He’s gone on to do some interesting things since then. And then I did the Diploma in computer science here in Cambridge, and have had a venue and a great course? The world’s first taught computing course started in 1953 did the diploma, and then have had a mixed career as a technology journalist, critic, commentator, software developer, consultant, man about town, whatever person about town. I currently sit inside the BBC research and development team. I’m not here on behalf of the BBC today. Nothing I say should be held to represent the BBC position on anything. I have to say that otherwise I’ll get into trouble. I’m here because I’m a mate of Mark. Okay? So you say mate of Mark, Bill Thompson said this, that’s fine, okay, I’m not BBC. Bill T. You actually have two accounts on X, Bill T and BBC Bill T that you can probably tell which one’s me, but if you can’t, it’s the one at the back with the curly hair and the naff jumper.

The person next to me, Mike was, in fact, a student here, and the person next to him also, Troy was a student here as well, and later Troy became my first wife and the mother of my children. So that’s the day I met her. You thought to have a photograph of such an epochal day in your life. If you get me for coffee later, I’ll tell you more stories about that day, because they’re good. However, I’m rabbiting on. Oh, the last thing to say then. So Troy and I met their psychological laboratory had children, if you watch the film, 28 days later, the opening scene where the rage virus escapes into the world, is set in that building. So when I watched it with my eight year old, first time, I said, that’s where I met your mum. OK, sorry, I am riveting.

I have a very straightforward approach to AI, which is, I hate the term, right? I refuse to say artificial intelligence, because I don’t like artificial and the word intelligence itself is corrupted. And anyone who’s read the psychological literature knows what a race based term is. I don’t think we should be using it. There’s no unitary measure of human, human capability. So artificial intelligence doesn’t work, and even the people who came up with it weren’t that happy with it. And the idea of computational rationality, which is what Stuart Russell prefers, seems to be a much more sensible term, but we lost that battle back in the 60s. So I talk about AI as it’s a term in itself. It’s not an abbreviation for anything, but that’s me being awkward.

Nonetheless, AI systems are designed to operate with varying levels of autonomy. Says the OECD. And I like this definition because it takes away all the anthropomorphism. It takes away all the sense, oh god. It might have a mind. It might have a soul. I personally, with my background in both philosophy and psychology and my experience in computer science, do not believe there is anything approaching self knowledge, if you want the short version, what we’re currently doing is implementing the early Wittgenstein’s picture theory of knowledge from the Tractatus, and what we should be doing is implementing the later Wittgenstein’s theory in philosophical investigations and trying to make machines that play or take part in the language game. And at the moment, they don’t again happy to explain all of that over coffee. The point is that what we’ve got instead, and what we really have is something very important, which is the creation of item, the computrainium.

Now, computrainium is a well established term of art referring to computational infrastructure. This is trainable computational infrastructure, and so it has some very different characteristics, but all we really doing is adding new capabilities to the computational matrix that underpins the modern world, underpins every business. It’s not fairy dust, it’s not magic, it’s not that special, but it is complex and rich, and it has emerging properties that we are not yet prepared for. So we need to think hard about it. And of course, it’s not just generative AI.

AI has been around as a term of art for 70 years. We’ve been deploying AI based systems. All of you are using them in your businesses all the time. This is some research from the BBC about extractive AI that we use to identify animals in the spring watch television series in the UK. And you know, that’s just good old fashioned AI, as opposed to this modern, new fangled AI, incredibly useful. The big difference with Gen AI, one of my colleagues says, is it’s randomness. It’s stochastic. The generative AI, in order to generate new stuff, whether it’s an image or a text or code or whatever, needs to have an element of randomness in it, it is fundamentally unpredictable. That’s its quality, and therefore, as a consequence, it’s fundamentally unpredictable, and that’s its problem, and there is no way around that, right? That’s like trying to solve that problem is legislating, say the value of pi is three, and then wondering why your calculations don’t work and your buildings fall down. It is at the heart of what these things do. But here’s the surprising thing, we are all random too, and yet we’ve managed to create this glorious infrastructure. So it’s not like randomness doesn’t mean you can’t do useful things, it just means you need to accept and work with it, and far too many of the claims made for Gen AI seem to be, oh, we can solve the problem of randomness. They are like politicians saying we can build secure back doors into your encryption systems. Just math harder. Just math harder, say the MPs. And we know it doesn’t work.

So again, we can talk about this later, but my belief so what we have at the moment is a world within which we have taught the machines how to think, and as Caliban said, our profit on is they know how to hollow out the middle class tax base. Okay, well, Hugh said, my profit on it is I know how to swear. But you know, it’s the same sort of thing that the machines do things which we think of as thinking, fine, but they be called thinking. I’m not that fussed about that word, but we have a big problem.

And the quote I didn’t put in here from Cory Doctorow is also a good one, which is,

We have not yet built machines that can do your job, but we have built machines that can persuade your manager to let them do a shitty version of your job.

And profound insight from from Dr, doctor, actually, he is a doctor. I was there when he got his honorary doc from the Open University. So check out Corey if you want to be made angry. So let’s worry about some aspects of this, because if we’re going to let these machines shape our environment, we need to do something about it. We need to think about the wider question about the information ecosystem in the age of AI. I do think it’s coming. Don’t think we’re going to avoid it. We just can decide how we develop these tools to serve people’s real needs.

And that was the starting point for my talk that I was asked to think about a couple of weeks ago for the UK AI conference about the world in 2107 so if the impact of AI is growing, if these tools are becoming more pervasive, what might we see now? Let’s see if this happens.

Yes, okay, you can’t hear it, but this is spot the tour guide. Somebody put a voice interface onto one of the spot robots, and it’s not freaky at all. They put a bowl of hat on as well and got it to do a little tour of its little space. Here is the charging station. Here are the blocks I like to walk over and stuff like that.

He’s fat. How do you like your job? Ah, Mr. Matt, my employment as a tour guide provides great satisfaction. (Video Audio)

Okay, now that’s just a bit off, frankly, but I’m sure they’ll be wandering around the corridors of the Fitzwilliam Museum within five years time. So we’re doing this sort of thing, both playful and otherwise. I’m a big fan of the Luddites. The Luddites from the 18th century in the UK, 19th century in the UK, have a bad reputation as people who destroyed the new technology of weaving machines and mills and stuff like that, and terrorized the nice capitalist owners of the factories who gave their workers such great conditions and produced such fantastic products, the reality as you may or may not know, but is somewhat different.

In the Luddites were a group of very skilled craft workers who jobs were being transformed by technology, who wanted to be in control of the deployment of that technology, and didn’t see why all the benefits should go to the owners of capital who built the machines and not the workers who operated them. And also noted that the machines at the time were producing yarn in particular, and cloth of poorer quality than was normally made, selling at a lower price, and it was just changing the world. And they didn’t object to the machines, in principle. They objected to the way those machines were embedded in the working system at the time, and they wanted to have some shape control over their own destiny. And they lost. They did not succeed in pushing back. And there is a question about what we might do around AI in the future, and whether there are parallels between that world and that world of 200 years ago, and this world, whether we can imagine humane deployment of AI and Gen AI into the world. And so that’s the framing.

So this is a say something I wrote a couple of weeks ago that I’ve fished out for you. We’ll see if you like it. So what is 2107 I’m looking back at today, talk about what happened, how we got here, stuff like that.

So greetings from 2107 you may not enjoy this bit, and I’m just going to check because it was a short talk, and so I actually wrote it down because I wanted to make some very specific points. So I’m going to move from here’s me waffling on in front of some slides to this.

The Letter

Greetings from 2107 I’m really glad to be here to share in your discussions. Let me tell you what it’s like here. Well, in the end, the global population bottomed out at about 4 billion. It feels like a victory. All our simulations showed that we came this close to total species loss. Okay, let me give you some context.

The term AI was invented 160 years ago in 1956. It’s eight decades for me since voice control systems emerged with things like Alexa and stuff like that. The LLM explosion was a couple of years, a couple of decades before I was born and when we look back at the recordings of that time, the excitement on everyone’s faces is just it’s delightful, it’s charming, it’s sweet. It was such optimism. However, not even your largest models could provide a path through the political complexity of working together to reverse the changes that were happening in the Earth’s climate.

It’s just 75 years now since global temperatures exceeded three degrees above pre industrial levels. We didn’t swerve the climate emergency. We went straight through it. We tried to remember enough to rebuild on higher ground, to grow plants in unyielding soil, to make things that could survive powerful winds. We learned to live without reliable electricity and sewage and connectivity or harvests. We explored what it means to be human, or what it means to be connected to the planet, instead of acting like it was a place to plunder before we moved on, and the surprising thing was man. The surprising thing was that we managed to get through it with only 4 billion fewer human beings.

Three decades of chaos, climate refugees lost, cities lost, growing land, water wars. It was hard. After the first few years, even the billionaires realized that Mars wasn’t going to happen, and so making life manageable, here was their only chance. But it took a long time before there was any sense that humanity as a whole might work together under a new dispensation.

It was around 2050 that things started to come together, and for the last half century, the world has taken a different path. Life’s much simpler. We find contentment away from our polished pebbles. Technology still underpins our society. We haven’t turned away from it, but it’s not extractive. It’s not operated in ways that make a very few people supremely rich and powerful or that support practices of oppression.

Now we certainly haven’t solved many of the social issues that face you all, but perhaps looking at the end of our species directly in the eye made us feel differently about small differences. Made us feel that we were more determined to resolve them, to work together. And that was sort of putting into context the small genetic differences or historic backgrounds we might have had to look beyond race and gender and sexuality. Religions remain, but we have a stronger sense of shared humanity and shared enterprise. And so religious differences matter less, and where they do matter, we look for understanding and accommodation. Lots of what we do is smaller, more grounded, more connected. We have remembered what we have remembered, and we built new patterns around it. We’re more patient, less acquisitive, more social. We don’t accumulate. There are no billionaires. We rework, remake, repair, recycle, limit new resource use. We have energy systems use wind and water and solar and nuclear. The systems that serve us are machines. We don’t imbue them with personalities or minds or souls.

In 2000, of the reasons why many startups got going and got going on the cheap was because the .com crash left a lot of infrastructure in the ground. Some of us may remember that. It turns out that the same thing happens when you seriously disrupt a weather system as things fell apart. It turned out it was possible to harden the tech infrastructure, even if we couldn’t save the food well. The h100 still worked, even though there was no Internet, Internet to connect them to for a decade. Old phones contain good chips. You can run a data center on local solar. Cold SSDs in a dead data center, they decay really slowly. You have to be very careful when you boot them up for the first time, it’s a bit like being in your command module on Apollo 13. But apart from that, it’s doable.

Rebooting the network was a lot harder. Most of the above ground infrastructure, like transmission towers for microwave hops or phone networks had gone. Tectonic activity took out a number of the big cables, and making new cables is really hard.

Today’s internet is slower. Lot of copper links, more of a mesh like the old times than the interconnected super networks that you may that you have today. It works. We design appropriately. UUCP has made a comeback. Joke for the nerds there.

So for around 30 years, from 2050 we basically plundered the debris of extractive capitalism. We’d left a lot of crap around frankly, thank you. We’re running quite a lot of solar, but making new panels is really. Hard, and we rely on the ones that survived mostly, and they break. We’re getting a lot of wind, but making turbines is hard too, and also the winds are much less predictable, so the turbines blow down a bit more than we’d like. However, it turns out, here’s a tip, a well designed small nuclear reactor isn’t much more complicated than a Victorian steam engine, and the tolerances are a lot greater than you’d expect. Me we lost a bit of territory, but not nearly as much as we feared. So it’s really the way going. Nuclear Engineering is a lot more straightforward than you’d think with our levels of technology innovation. So do give it a go.

And once we’d re established, reliable electricity and a working network, we started to think about how to make the advanced stuff again, retooling to be able to manufacture things like GPUs. We had all the old data, the old models, the old apps and services and code repositories to be mined, even some of the old machines. Now we’ll probably never get to two nanometer chips, but what we’ve got serves our needs, because we’re not actually in the same race.

We weren’t starting from scratch. We were standing in a different place, a place where the questions were not about getting rich, but about making a sustainable world that reduced starvation and thirst and fear and pain and suffering felt like the right thing to do, and building machines to support and sustain a bearable life for everyone is it turns out, a lot easier than satisfying the peculiar needs of already privileged groups.

We don’t build humanoid robots. They were mostly a vanity thing anyway, but machines that amplify people’s capabilities. We aren’t looking to replace human labor, because we know that human labor brings value to process and benefit to the person, if it’s done in service of something other than the ego of a desperate billionaire. We aren’t searching for AGI anymore, or for augmented humanity, augmented humanity, but an augmented society. And I think we’re doing okay, not everywhere, not everyone, but a good enough proportion to claim we’re on a sustainable path. More in more in harmony with the world, more in harmony with the world we’ve remembered that we’re part of. To borrow an analogy from some old friends, the dark mountain is dotted with lights.

However, and this is one reason why I’m here. We are not clear yet. CO2 levels aren’t rising, but they aren’t falling fast enough, and we don’t know why our models are not good enough to figure it out. This planet is still pretty mysterious to us. We’re still seeing the temperature rising and weather systems are getting disrupted. We lost a significant chunk of Madagascar five years ago. There’s no sign that Australia will be coming back any time. Florida is the new dog a bank, If only, if only we had seriously tried to put the AI systems you have today at the service of humanity, instead of seeking a silicon Godhead or imagine that a tool to design better PowerPoint decks would save us.

Conclusion

In the end, we didn’t need the machines to write our code for us. We needed people to talk to each other, recognize that we have to work together to solve our common problems. But maybe there’s something in what you’re going to say today that will help us over the next few years. Because, in fact, now come to the hard bit. That’s the point of me being here. So because I have to tell you something now that you might find rather difficult to deal with, you’re here to share your insights as part of a simulation, although I presented this as some theatrical version of time travel. Time Travel doesn’t exist. What I’ve done is to spin up a simulacrum of BoS 2024, as a probe into our future options, trying to figure out why you all made some of the choices you made 40 years. Trying to figure out some of the choices you made and why you made them. We’re looking to potential, a potential danger to the food web again.

So I was asked to explore how researchers and business people in 2024 in the first quarter of the last century dealt with these problems. And so we’ve created this simulation. And I think you’ll agree it’s quite realistic to do that. Now we had to use 2024 because it’s the last year we have usable information for from about this point, the internet is so full of generative AI garbage that you’d all have six fingers and be zombies. So you’re basically the last credible humans we can spin up. And we really do appreciate you being here. Obviously the sim will be closed down at the end of the day, so I’d just like to thank you all for all of your insights and everything you’ve given us, and to remind you that do feel free to eat all the cake and pastries because it won’t have any long term effect. Thank you.

Q&A

Mark Littlewood 

Amazing Right. So, I know people are going to have questions, but I think you should give them questions, yes.

Bill Thompson 

So that’s the point. As I say, I’m here, and I do think quite mean, I think there’s quite a good little model of me, frankly, and the rest of you look really realistic. I mean, there’s some, there’s some good spark coming from those glasses. I will congratulate our engineers when I get back for my coffee later. And you know, obviously you all won’t be here not being real. And so I really could appreciate some guidance, right. Here we are in 2107 we’ve hit the climate crisis because you didn’t do anything about it. I’m interested in ways you all think we might have done it better from now going on, given what you know about today’s world, the tools you have available, the capabilities of AI. I’m interested in any insights and help that I can take back to my future, well, my present actually from you all. You are the best that our technology can make. So see what you got. There’s microphone coming to you.

Audience Member 

So I have one observation. So my children are very into the environment, and my daughter won’t fly, and she’s very into getting clothes from thrift shops. And what I see is a very big differentiation and a polarization within our society. Because certainly within her generation, there are people who are buying fast fashion, and you wear it one day a week, and then there are people who are very frustrated and trying to do everything they can to save the planet. So I think to say we didn’t try anything, I think, is a bit you have all groups. You took us to a to a Greta Thunberg speech in the South Bank. And outside one door, they had just stop oil, complaining that Brett wasn’t doing enough. And outside the other door, it was piers, COVID and his mates saying what climate change. So I think part of our problem is, is we haven’t got agreement on what the issue is, on what the critical issue to us is. So I take issue with your point that we aren’t doing it. I think some of us are desperately trying to do it, and some of us are burying our heads in the sand.

That’s an important point. And thank you for that corrective. I think, looking at those who were trying hard and then thinking, how could we amplify what they’re doing, is it’s quite a useful thing, thing to note, particularly at the moment, because we do have people in my time, who are who think that things will be okay, that, yeah, we’ve gone through the big one, and we’re going through a smaller one, and some of us think it’s going to be a big one again, and we might actually lose the species. So building on the voices of those who have already found a fact, who have already stated their understanding of the situation is a really good point. Okay, so you’re right.

Bill Thompson 

So in fact, what we might do is, when we reboot this, because I think you’re about number 473 version of this, when we reboot it, we might bring in more people. We might up the Greta Thunberg scale and see what we get next time. So thank you for that one. All right. Okay, so can’t see you. And then you perhaps.

Audience Member 

So from the things you’re talking about, I presume you’ve read surviving the future by David Fleming. Yeah, so that talks a lot about trying to reintegrate a lot of the social things that we’ve kind of forgotten about, and having less less obsession with money. And I think the the way our AI is going at the moment, there’s going to be a lot less jobs. So I think we need to support, we need to start thinking more about how we’re going to support those people who haven’t got jobs, and how we’re going to help them to find useful meaning in their lives. And maybe, if, maybe, if we can, maybe, if we can, when we can do that, and turn all that energy that people have got that’s not going to be generating cash through a job as we know it today, how we can return that kind of energy back into our society, helping us to reduce the the things mean we need, the stuff we need, and getting more out of the relationships we have of as people. I think if we can get ourselves more on that track, maybe we can start reducing we have less need for as much energy and as much as much of all these, all these climate change drivers that we have.

Bill Thompson 

I think that’s very useful, and that sense that we could slow down the flywheel. And also, instead of thinking, what jobs are the people whose careers will be derailed by generative AI, the lawyers, the middle managers, the data scientists, the consultants, the software engineers, none of you, but  what new jobs do they have instead? Okay, how does society work without the need for those people to be part of a money economy in the same way and then just slow down that speed that could, that could have helped us avoid getting to where we are, and it’s also, I mean, we’re definitely have moved on beyond that to a large extent, but it’s useful to take that on board. Thank you.

Audience Member 

Thank you so much. It was really nice to understand different perspectives. I have never been, like, into climate change or such things. I like, always love to be. Just think about the software and Gen AI, as you mentioned, is random, but has positive and negative as well. But I really like one line which you mentioned, AI at service of humanity. So it’s AI is definitely, is the future of software. But when we say humanity, can you just help me to elaborate how we can think in that direction,

Bill Thompson 

Augmenting humanity. Yeah, so from our perspective, there’s a lot of discussion about things like existential risk and AI and artificial general intelligence, and that’s never felt helpful, because it focuses on an attempt to emulate the way human beings operate, and normally take the space of a human being in the world. And we think that the actual role of any technology is to assist and support people in their individual lives, in their social lives, in their business work and stuff like that, so to augment what we are capable of and to offer support, not to attempt, perhaps, to do some jobs that we do, but not to replace the role of a human. So keeping humans right at the center of things. So from our perspective, that becomes much more important than the quest to create a machine that can think like a human does. So it’s that different orientation I was referring to.

Audience Member

Whatever Gen AI is generating. It’s whether it’s right or wrong, fake or real, in a way we have something to help or to give a sense of something to get, to get work done easy for humans, isn’t that?

Bill Thompson

Yeah. Oh, okay, so yes, if, if the system, if the machine, if the Gen AI, can fill a space within, let’s say, a business that is a value to the business overall, and which helps support and sustain the people who are the business, that’s great. It’s when you start trying to put it in as a replacement that what we found looking back from where I am, that’s when things started to go wrong, because it started to unpick the social fabric in unhelpful ways. It started to lead to the sort of problems of people losing their jobs and having nothing to do and stopping valuing themselves, and it distracted in our timeline from the attempt to come together to resolve the issues of climate change, which we see now looking back as the most important problem facing you.

Audience Member 

But this happened with every machine revolution. Like earlier, people are working manually. Then machine came in textile or like in every industry.

Bill Thompson 

Okay, I’d like other people to have a chance to talk, but, yes, but I would say that what we found is this one was different. Okay, so, and we can talk over coffee, perhaps, about how it’s different, but you’re right that we’ve been through this process. We think this one accelerated.

Audience Member 

Thank you. Thank you. I’m both more pessimistic than you and also more optimistic. My pessimism is there’s absolutely no way that enough economies and populations will cooperate in a timescale required to bring about the change necessary to avoid catastrophic consequences. And I also don’t believe that new future you outlined is possible, because when 2 billion people face extinction, they’re not going to take it lying down, and therefore the revolt in the world is going to be horrendous. And therefore, my optimism only comes from our human creativity. We’re going to have to use technology to profoundly reinvent how we do things in ways we cannot currently do in order to somehow get ahead of the problem and avoid what Thomas Malthus once anticipated. We’re going to have to do it again, and I don’t know how, I mean, we can throw out quantum computing and fusion and this, that and the other, but if we don’t have something, I don’t think we’re going to, in a cooperatively way, cooperative fashion, solve the problem.

Bill Thompson 

Well, from 2107 I can tell you that we didn’t. As a species, we didn’t come together. There was no technological deus ex machina. We didn’t solve the problem. We didn’t swerve. We went straight through. And from all the records, it was awful, like really awful. And we did come very close to losing the species. You know, there was only a few lucky bits, you know, basically the the change didn’t happen quite fast enough to take out our last readouts. We built down just enough. The infrastructure was hardened just enough. The approach to climate refugees was not quite cruel enough. It was all just, just enough to avoid it. And from my perspective, you know, yeah, if you could come up with a solution in the last 10 years, then you might be great. Coming up with a solution in the next 10 years is going to be tricky. And if you are in any sense involved in in the world as a business leader, whatever, from my perspective, not taking account of what is happening and the changes are happening, is going to be quite damaging to your business prospects and to your bottom line in 10 years time, because your bottom line may be underwater, as it were. So, yeah, I share your pessimism. I’d love that. Sorry. I think Mark is choosing people. Are you choosing?.

Mark Littlewood 

Yeah, John, at the back. Then, Adam.

Bill Thompson 

Someone’s got a microphone. Okay, but yes, so thank you for that. And yeah, let’s have a last cup of coffee later.

Audience Member 

Do you think we’ve missed the puck? There’s a lot of focus on AI and technology and looking at regulation to control it and make sure that it’s kind of is used in the same way that something like IVF could have been incredibly destructive society, but was loss of regulation allow that kind of innovation to happen. It sounds like the problems we have are social problems, not technological problems, and the impact of social media on society over the last 1015, years, have we missed an opportunity there to provide a framework and guardrails for safe social media use that allows us to come together as a species, rather than create the division.

Bill Thompson 

Oh, that’s very thoughtful. I like your reference to IVF. So invitro fertilization, particularly in the UK, when that technology was being developed. A sensible government at the time realized that it needed to be regulated because it was about to do things with human embryos, and so a whole authority was created to look after academic research, industrial research and deployment, and is generally deemed to have been very successful in terms of accelerating the rate of responsible research, allowing funding to come in, and actually creating a situation within which we are able to do good science and help people that be a reasonable framing of it.

Audience Member 

Yes, thank you.

Bill Thompson 

No. And I heard somebody talking about survey of AI recently, and they were talking about previous attempts to control other technology, and they said, Well, maybe the atomic bomb. And they had no one, they had no knowledge, no historic knowledge, of what we done with IVF in the UK. And I think it’s a brilliant analogy. There are a number of initiatives to control AI, the UN is going to launch its guidance AI safety councils publishing in June or July. We’ve got the EU AI act, and all this is happening, but that framing about maybe we are incapable of allowing that regulation to be discussed and developed and then deployed in a reasonable way because of the damage that social media has done to the public discourse is one I’m going to have to think about, but it’s, it has that ring of, Oh, something going on there.

I remember Charlie Stross, a science fiction writer, said the invention of Twitter in particular was like giving us all telepathy. And it was that damaging and destructive, because instead of looking at a room like this and thinking, what a nice bunch of people. I now know that most of you, in fact, pretty odious, and I wouldn’t want to spend time with you, and I didn’t know that before, and I could work on the soft assumption that we could get on. And now I’ve seen what you write. I know what you’re thinking. I’ve looked and that was just not on right? And there’s so that all these reasons not. To like each other have been amplified, and those of us who’ve had quite a positive social media experience did so because we live in a bubble of privilege and were lucky, but for many people, it’s been completely toxic. So that sense that we’ve broken the space within which we could have a sensible discourse to help us cope with this technology something I will take away from this. Thank you. That’s really helpful.

In particular, in my time, we’re seeing re emergent social media with poor moderation, so we might need to do something about that. There’s somebody at the back waving their hands, so at some point, there needs to be a microphone to come to you.

Audience Member 

I know this is a simulation, so obviously the answer that you give, I’ll try and persist for the next version.

Bill Thompson 

Yeah, and by the way, I’m really sorry about the t shirt we gave you in the sim. Yeah, those chunky gifts just were on a drive somewhere.

Audience Member 

Maybe another collar in the next boot. I was just curious. So in the 2107 what’s been the impact of the improvement in quality of life that I hope there was, and potentially longevity as well. How has that affected? You know, the society and the numbers of people on the planet.

Bill Thompson 

We don’t focus that much on longevity, so much as living a good life until you die. So it’s not about extending life. Is about.

Audience Member 

The quality of life.

Bill Thompson 

Exactly. It’s the quality of life. So, you know, most people will die in their 70s or something, but nothing too horrible. Again, it’s the focus of medical science is not on providing something which could then be sold for billions of dollars to a few rich people who are desperate to sustain themselves for over a century, but rather to make things work for everyone. And so what we’re finding is, because people’s active life is extended, they are able to take a greater role in society, and we value them more. My friend Naomi Alderman, the writer, pointed out that the invention of writing was really bad for old people. So until the invention of writing, you needed your elders to tell you whether the river ever came up to that mark and what that sign on the mountain meant. And once you could write it down, they could die because you’d written it down, that knowledge had been captured and they were the elders were no longer as useful to you. And I thought that’s quite an interesting insight. As we pull out all of our accumulated knowledge of humanity into our large language models, whose knowledge becomes less useful when it’s embedded in their consciousness, ask yourself that question and begin to worry. So really, what we do now is try to value everyone, and what medical science we have is mostly around quality of life rather than heroic life extension.

Mark Littlewood 

Thank you. Adi, your time has come.

Audience Member 

Okay. Hello, so sorry. I got a little late because I was stuck in 2200 and trying to make it back here. So God.

Bill Thompson 

Does that mean I’m in your sim?

Audience Member 

And we did, we did hit the reset button there, so we were not as fast as you.

But I’m just like, curious about, if we’re talking about augmenting capabilities, what is it that we did, or that we could do that would make 2007 better when it comes to rapid resource reforest, reforestation, reforestation. I mean, we have this big sand in the Sahara Desert. You know that can become like a big regulator. And I know that some African city to Mark’s comment earlier took on their own to start like a green belt from like, you know, Western Africa to Eastern Africa. They just did not have like the means to make those type of like projects happen, but with augmented like capabilities that AI could give us, what could we gain from like, those type of ecological efforts and like, make it profitable for the population that are doing it in The first place? And then my quick second point is, can we all quit on fighting and bashing on social media and understand that maybe we are just the one who are not using it right? Because if we look at challenges we had, like many challenges that were that were, that were done and that was very successful. How about community challenges?

What about if we said, like to the people of, I don’t know, Kent or wherever, we have a challenge for you to make, like, your soil, more nutrition, and I don’t know, but we could use the power of social media on, like, getting us the right type of attention, and then, like, using social media for good. I just feel like. Just don’t try, like, hard enough yet, and don’t understand the power it has to get everybody, like, involved, from Timbuktu to like, you know, to Australia in, like, in matter of minutes, we really undermining, like, the power of that. Thank you.

Bill Thompson 

Okay, so those are both very good points. To the social media point. There’s a lot of interesting work going on, led by people people like Ethan Zuckerman in the States, around small communities and the use of social media with clear boundaries, smallish communities to do things and that is successful. The issue seems to be a billion person timeline is what causes the problem after the jackpot, when things fell apart, we only had small communities using social media locally, and that was successful. Probably facing 2107 is we’re starting to link them up again, and finding the same toxic dynamic is starting to evolve. So how do we moderate and regulate to get the good i think is a good question. But to your substantive question, and I’ll finish now, because I’m conscious of the time we did it right. So there were three elements. Firstly, we had we hadn’t forgotten everything. So we had CRISPR technologies to re engineer some of the tree species, to make them much more robust in the soil conditions and weather conditions that we had. We made them smaller and harder so they could withstand the wind. So we had the trees. Second thing is we had we understood the value of human labor, so we didn’t try to build machines to go and plant a million trees. We got a million people to plant a tree, and that gave them meaning in their life. That was away from being a job and was valuable, and also the planting of the tree connected them to a future they wouldn’t see, because that future was 1020, 30 years ago. And so using the technologies we had to hand the people who wanted to help, we gave them a vision and a connection to a world that was beyond their imagination, and that was what transformed things. That was one of the things that made the difference. Then those trees, now 3040, years later, are mature. They’re stabilizing the environment more. They’re one of the things that give us hope. So everyone go and plant some trees, even though you’ll never sit under their shade? Thank you.

Bill Thompson

Technology Critic

Bill has been working in, on and around the Internet since 1984, thinking, writing and speaking about how digital technologies are changing our world and how the network can be a force for good. During the 1990’s he was Internet Ambassador for PIPEX, the UK’s first commercial ISP, where he developed websites for Comic Relief, the Edinburgh Fringe and Anne Campbell MP before moving to Guardian Newspapers as head of new media. He established the paper’s first website in 1994 and was responsible for many online projects, including Eurosoccer.com in 1996.

Bill currently leads Future Value Research in BBC Research & Development. He is also well-known as a technology journalist and advisor to arts and cultural organisations on matters related to digital technologies. From January 2001 to April 2023, he was the regular studio expert on the BBC World Service technology programme Digital Planet (aka Go Digital and Click). He still appears on air as an independent commentator. He is an Adjunct Professor at Southampton University and a member of the board of the Web Science Trust. He was awarded an Honorary Doctor of Arts by Anglia Ruskin University in 2016.

Find out more about Bill

Next Events