In short
Neil Lawrence (DeepMind Professor of Machine Learning, Cambridge) argues AI’s biggest danger is already “socio-technical,” not Terminator-style; fast machine decisions can destabilize society (e.g., flash crashes). He contrasts human vs machine intelligence via information bandwidth, discusses ChatGPT/Turing-test limits, and emphasizes empowering people with control. He also covers uncertainty in ML (his thesis), real-world deployments (F1 strategy, Amazon supply chain), and societal governance (digital oligarchy, data privacy, data trusts).
Guest background
Mechanical engineer (Southampton), oil-rig field engineer (North Sea/Irish Sea), PhD on uncertainty in neural networks (Aston/Cambridge), research at Microsoft, professor at Sheffield (Gaussian processes), co-founded AI startup acquired by Amazon; founded Data Science Africa workshops; now Cambridge professor.
Key claims
ML should be used to influence local “front pages,” not chase journal covers; autonomous tech requires social adaptation; data foundations must be fixed to avoid scandals.
Notable examples
AlphaGo “Sedolian voids,” motion capture for Godfather game (Marlon Brando face), sequencing/transcriptomics, Data Science Africa (health/agriculture), Ferrari F1 strategy using Gaussian processes, Amazon AWS/Alexa/supply chain, Cambridge Analytica, flash crashes.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Chapters
Tap a time to open that second in VOAI: Threat or Tool?
2:21 to 2:56
Neil discusses the misconceptions surrounding AI as an existential threat.
“Now, as I alluded to in the introduction, there are voices, you know, including colleagues in your own field who think AI represents a serious existential threat to humanity.”
Understanding Machine Intelligence
2:56 to 3:40
Explore the differences between human and machine intelligence.
“So when we say existential threat, computers, AI, maybe one day will be as intelligent as humans, but it's a long way off.”
The Turing Test and AI Capabilities
3:40 to 6:05
Neil reflects on the implications of AI passing the Turing test.
“And it's these effects often that are the ones we should be paying attention to.”
Driverless Cars: The AI Challenge
6:05 to 7:39
A discussion on the complexities and social contracts of driverless technology.
“It could bring people closer to the machine, give them more control over what the machine's doing.”
From Oil Rigs to AI Research
7:39 to 9:22
Neil’s journey from mechanical engineering to AI and neural networks.
“University of Southampton to study mechanical engineering.”
Challenges in Machine Learning
9:22 to 11:12
Understanding uncertainty in machine learning models based on Neil's thesis.
“So maybe you can explain what these are.”
Applications of Gaussian Processes
11:12 to 13:55
Neil explains the fascinating applications of his work in animation and healthcare.
“and then you have a ball coming in the top and the position it's dropped from is like the input location, could be to the left or to the right.”
The Emergence of Genomic Sequencing
14:01 to 15:10
Explore the development of genomic sequencing and transcriptomics and their impact on health research.
“So what sort of things are we talking about there?”
Data Science Africa and Local Innovations
16:09 to 18:19
Discussion on Data Science Africa's role in promoting machine learning across Africa.
“a network that runs machine learning workshops across the African continent.”
Challenges of Technology in Africa
18:19 to 19:53
Insights into the managerial challenges faced in sustaining technology solutions in Africa.
“So from the big tech companies perspective, this isn't a challenge.”
Show all 16 chapters
Neil's Consultancy with Ferrari F1
19:53 to 21:51
Neil shares his experience consulting for the Ferrari F1 team and applying data strategies.
“opportunity for a car lover like you you're asked to be a consultant for the ferrari f1 team Yeah, that was amazing.”
Transition to Amazon and Lessons Learned
21:51 to 23:54
Neil discusses his move to Amazon, focusing on machine learning applications and challenges.
“I hear when you moved to Cambridge, your wife gently suggested it was time to get rid of a prize collection you'd had for years of various old computers.”
The Digital Oligarchy and Data Control
23:54 to 25:11
Exploration of the concept of digital oligarchy and implications for data control.
“Earlier, you spoke about the disproportionate and unhealthy control of data by the digital oligarchy of big tech.”
Returning to Academia
25:11 to 27:11
Neil reflects on his return to academia and his work on practical applications of machine learning.
“you made the move back to academia, to Cambridge University, where your dream job came up, the inaugural deep mind professor of machine learning.”
Data Trusts Initiative and Future of AI
27:11 to 28:00
Discussion about the Data Trusts Initiative and concerns about AI's impact on society.
“But there's still a lot of confusion and even mistrust around AI.”
The Importance of Communication
28:00 to 28:26
Explore the role of communication in shaping our interactions and responsibilities.
“It's part of how we communicate and how we move forward.”
Transcript
Automatic transcript. May contain errors.0:00Neil Lawrence:This BBC podcast is supported by ads outside the UK. Pop quiz. What's in your kid's lunchbox? At Whole Foods Market, they've already done the studying. Over 300 food ingredients are banned from their shelves. No hydronated fats in the peanut butter and no high fructose corn syrup in the cookies. And for sandwiches, there are no synthetic nitrates or nitrites in any of their deli meat. So you can pack lunchboxes with peace of mind. Get back to school ready at Whole Foods Market.
0:59Neil Lawrence:Where your story starts.
1:04Hello.
1:05Neil Lawrence:Today it's become obvious that artificial intelligence, AI, is going to play a huge part in our future. But not everyone's entirely clear or indeed in agreement on how that will pan out. And a lot of conversations around AI and machine learning still reference the idea of a dystopian future where Terminator-style robots are our overlords. But don't panic. My guest today says that's all rubbish. Neil Lawrence is the DeepMind Professor of Machine Learning at the University of Cambridge. A mechanical engineer by training, Neil's story is one of contrasts. He's worked for both academia and Amazon.
1:42Neil Lawrence:He's plied his trade on oil rigs and helped set up data workshops in Africa. Perhaps most and probably, he's both a keen cyclist and a petrolhead. During his career, Neil's also been involved with deploying AI and machine learning across fields as varied as movie animation, Formula One strategy and local planning applications. And he says ultimately all his efforts are about making a difference to our everyday lives, explaining, as scientists and academics, we do well to focus less on getting the front cover of scientific journals and more on influencing the front cover of the local paper. Professor Neil Lawrence, welcome to Life Scientific.
2:21The Life Scientific:Thanks very much for having me, Jim.
2:23Neil Lawrence:Now, as I alluded to in the introduction, there are voices, you know, including colleagues in your own field who think AI represents a serious existential threat to humanity. But you don't agree.
2:34The Life Scientific:Well, I think there's a danger of what I would say is a socio-technical existential threat. But I think we're already in the middle of that. Digital tools are already isolating professionals from their decision making. And my worry is by overly focusing on these futures that are a long way away, we're ignoring the challenges society is facing today that could be made worse or better from this technology.
2:56Neil Lawrence:So when we say existential threat, computers, AI, maybe one day will be as intelligent as humans, but it's a long way off.
3:04The Life Scientific:Well, the nature of their intelligence is very different. In some sense, they've already exceeded us across many, many parameters. So just look at the stock exchange where computers are used to make trading decisions at speeds that are inconceivable for humans. And what did that drive? It drives flash crashes. There have to be systems in place to prevent the entire financial system crashing when these computers get a bit carried away. So that's not because they've become intelligent in the way that we're intelligent. But technology through that type of effect, fast decision making, can already have a dramatic effect on society.
3:40The Life Scientific:And it's these effects often that are the ones we should be paying attention to.
3:44Neil Lawrence:And so we're confusing different kinds of intelligence. The speed of communication, of computing, doesn't necessarily mean thinking.
3:51The Life Scientific:Precisely. And the key difference between a human intelligence and a machine intelligence is the rate of access to information we have when we're communicating with each other. So I'm talking to you now. We're sharing information around 2 ,000 bits per minute. It sounds quite a lot, but two machines communicate at 600 billion bits per minute. But it doesn't make the machine more intelligent than us because, you know, it's more what you do with this information access. And so much of the interesting things about human intelligence are how we manage this narrow bandwidth, how despite this limitation, we achieve extraordinary things.
4:26Neil Lawrence:What's your take then on the recent news that the latest version of ChatGPT has passed the Turing test? That is to say, it fooled the majority of testers in a scientific experiment into thinking that it was human.
4:37The Life Scientific:It's interesting because you sort of have to go back to an argument between Turing and colleagues that led to the Turing test, this notion about whether it made sense to think about a computer writing a sonnet. And what Turing was trying to show was the notion of thinking didn't really make sense as a scientific concept. He was saying, look, I could build a machine that could fool you. Is that thinking or not? And of course, machines have been able to write sonnets for a few years now, very good ones. But I think that the underlying point is that the origin of the sonnet is important. So when the machine's writing a sonnet, it's doing it based on all human written work and the sort of, in inverted commas, understanding of humans that comes from reading everything we've written about ourselves.
5:24The Life Scientific:It can't feel the sonnet. It can't sit there and have the hairs on the back of its neck go up. in the same way that a human does. So even if you can't tell whether something was human-written or not, there's still a significant difference between the origin of that feeling, the origin of that thought. Not that it's not interesting and exciting that machines do this, but it's not the same. Well, clearly there are lots of exciting prospects here, but presumably there are risks as well? Extraordinary risks. And we're already seeing those risks panning out with previous deployments of machines. But I think the extraordinarily exciting thing about this new wave of machine learning technology is it could also make it better.
6:05The Life Scientific:It could bring people closer to the machine, give them more control over what the machine's doing. So you can imagine a world where we're asking the computer to do the thing we need rather than having to adapt to the thing that it can do.
6:18Neil Lawrence:So rather than it being controlling our lives, we're empowered to use it the way we want. Now, I've also heard that you're a car fanatic, Neil. So what's your take on one of the main example applications of AI, driverless vehicles?
6:34The Life Scientific:I mean, do they hold much appeal to you? So I suppose I was really a car fanatic in the 1970s. I bought an MGB GT as soon as I could afford one that I had at university and still have today. But when I look at driverless cars, I think that's an extraordinary technology. But it's been clear from the start that the great AI fallacy holds we can't just deploy them on the streets and expect them to assimilate each other one of the ways i often think about it my wife's from naples and and if you ever need to cross the road in naples the first thing you do you have to make eye contact with a driver right now the amazing thing about naples and one of the things you learn is when you make eye contact with a driver you're entering into a sort of social contract with the driver that if you now walk out because that's what you do is you make eye contact and walk out yeah that if you walk out they're not going to kill you now if you could guarantee all cars would stop if they were autonomous vehicles then we could all just walk out in the street in front of them cause traffic chaos so yes these technologies can emerge but they typically involve us adapting to the technology in some way and and restricting our pre-existing freedoms interesting so 1991 neil you went to
7:46Neil Lawrence:University of Southampton to study mechanical engineering. And after graduating, you got a job as a field engineer on oil rigs in the North Sea. That's an interesting choice.
7:57The Life Scientific:It was a choice that was driven by the fact that unfortunately, the type of role that I wanted to take when I left in terms of creating and building things, it just seemed to be almost non-existent, certainly at the sort of salary level, you would expect to be a professional. And it seemed to be a job with a lot of adventure and a job where you were getting up close with some very interesting equipment. What sort of work were you doing for them? So I was based in Great Yarmouth and then I had to get called out to oil rigs in the Irish Sea when one of two things happened. Either they'd finished drilling the well and they wanted to know where the oil was so they'd expose the rock formation.
8:37The Life Scientific:And I was measuring that with tools that measure density, electrical resistivity, sonic transit time these different properties or they got their drilling rig stuck a bit like when you get your drill stuck in a wall and i had to go out and fire off explosives near the stuck point so that they could free it off and that used to take hours and then i would get about two or three hours sleep and then they would get stuck again and then i'd have to go through the whole process again did you enjoy the work in part it was fascinating work the diversity of people you met and worked with, and of course, the knowledge people had.
9:12Neil Lawrence:Well, after a couple of years on the rigs, you decided that you'd quite like to go back and do a PhD, specifically in an area you'd become interested in called neural networks. So maybe you can explain what these are.
9:24The Life Scientific:Well, actually, I'd been on the rigs and I was reading New Scientist. I read about these neural networks and they were being held up as an example of something that could learn from data, that instead of deriving the rules of logic or having to go through all the underlying science, You could just take data, a bit like humans do, and then you could use that data to make the decision directly.
9:43Neil Lawrence:And the name neural networks comes from neurons. It's the way the brain works, interconnected neurons. There are many of them, billions of neurons in our brain, but many more connections that join them up. And that makes it a very powerful way of processing data, doesn't it?
9:58The Life Scientific:That's right. And today, as you say, we're getting neural networks with sort of many billions of parameters. And what's extraordinary is scaling up this simple idea to enormous quantities of data, but leveraging the 300 million times faster nature of the machine. That's what's giving us these performance. And it's sort of unimaginable data, more data than any human could ever conceive of seeing in their lives. So although it's initially inspired by human intelligence, the way it's getting there is utterly different to anything you and I are doing.
10:27Neil Lawrence:Well, in 1996, Neil, you started your PhD in the Neural Computing Research Group at Aston University in Birmingham. But your supervisor moved to take up a new job and you ended up following him and completing your PhD in the computer lab at the University of Cambridge. Tell me about your thesis, because it's a topic that's still hugely relevant in terms of how we deploy machine learning today.
10:50The Life Scientific:So the thesis focused on uncertainty. So one of the problems in computer science and the problems with these methods still today is that really intelligent behaviour emerges when you're dealing with what you don't know. So trying to get these models to integrate uncertainty was a big challenge and still is. Now, I sometimes think of it as like a pinball machine, a traditional pinball machine with rows of nails. and then you have a ball coming in the top and the position it's dropped from is like the input location, could be to the left or to the right. It falls down the machine through each layer of nails and then eventually it emerges at the bottom.
11:29The Life Scientific:Now you can imagine that by moving these nails around, you could guide the ball to start from different places and end in different places. And in effect, that's what we're doing with the neural network. We're throwing a lot of balls through the machine, seeing how they drop and moving the pins around, adapting the parameters to do that. But it's not one dimension in, one dimension out. It's millions of dimensions in potentially and maybe thousands of dimensions out. So it's a sort of pinball machine in some enormous space. So what goes wrong with that is it turns out that there's regions that the balls never touch.
12:06The Life Scientific:So there's pins that are just in these locations because that's where they were at the start. And if the ball hits those regions, then the machine doesn't behave well. And this happened to AlphaGo when it was playing Lee Sedol. He managed to play the machine into one of these regions and it made a stupid mistake. I find that extraordinary. So I think of these regions as Sedolian voids. And the way we try and deal with those regions is just saying, well, until we've seen that the ball's gone in there, we'll just be really uncertain about what those pin locations are. and this was the basic idea that I was working on in my thesis and that sort of notion of needing to use and exploit the uncertainty still underpins almost all my work today.
12:50Neil Lawrence:Well after your PhD and a subsequent research year at Microsoft you took up a lectureship at the University of Sheffield and within a few years you were starting to make an impression specifically in what's known as Gaussian processes. Basically what you're doing is a statistical model that helps you make better predictions. Now, your work had some very interesting applications, for example, in animation.
13:12The Life Scientific:Absolutely. Yeah. So my particular innovations allowed something that's almost identical to modern generative AI, but it could use very small quantities of data. So we ended up working with movie companies and games developers to support them in motion capture. Motion capture is sort of the pins around a body that allows us to animate Golem in Lord of the Rings. In fact, Golem was a little bit after this. So this was to animate a face and to be able to take that face and project it onto a creature. And one example I remember was there was a Godfather game that they were using these techniques to try and animate the face of Marlon Brando, who, of course, is no longer with us.
13:49The Life Scientific:And it was incredibly exciting to see these techniques, these ideas being deployed by people at the state of the art.
13:54Neil Lawrence:Very interesting. Well, also around this time, you became interested in biological applications of the technology. So what sort of things are we talking about there?
14:03The Life Scientific:Well, just at that time, at around about the turn of the millennium, we had the emergence of sequencing and the first sequencing of the human genome. But alongside that, I think less well known about, we also started developing techniques that allowed us to measure which genes were being produced in which cells, something called transcriptomics. and thousands of genes in every cell. You could measure them. And this scale of data was very new and required new approaches for these extraordinary new questions we could start asking. And it became a real focus for you, this intersect with health and medicine.
14:39The Life Scientific:It's something that drove my research for the best part of 15 years. And what was so wonderful about it was because they were competing with a private company, the people who were at the Sanger Centre trying to create the genome had to share across academia their data. And this built a culture of sharing data and sharing ideas and algorithms that meant that this community was actually at the heart of some of the most interesting machine learning. And it's still true today.
15:09Neil Lawrence:Pop quiz. What's in your kid's lunchbox? At Whole Foods Market, they've already done the studying. Over 300 food ingredients are banned from their shelves. No hydronated fats in the peanut butter and no high fructose corn syrup in the cookies. And for sandwiches, there are no synthetic nitrates or nitrites in any of their deli meat. So you can pack lunch boxes with peace of mind. Get back to school ready at Whole Foods Market. Apple Vacations, where your story starts. The Summer Savings Event is here at Apple Vacations. It's the perfect time to bring everyone together for the getaway you've been dreaming about.
15:45Neil Lawrence:Book between July 24th and August 27th and enjoy up to 55 % instant savings. From relaxing beach days to unforgettable moments with the people who matter most, your next vacation starts here. Travel through July 31st, 2027. Start your vacation today at AppleVacations.com or through your travel advisor. Apple Vacations, where your story starts.
16:08Neil Lawrence:Well, Neil, it was also around this time that you got involved in setting up Data Science Africa, a network that runs machine learning workshops across the African continent. How did that come about?
16:19The Life Scientific:I had colleagues, in particular a guy called John Quinn, who had moved after doing his PhD in a similar area to mine to Kampala and had started working there to see if these technologies were useful in the African context. And I was just fascinated about the things they were doing around the availability of 3G, then 4G networks for mobile phones. So you go from a situation where there's no landlines, not very good roads, certainly no canals and not many railways, to a situation where everyone has 4G. That makes society radically different. And as long as you can educate people locally and support them in delivering the type of solutions that are tailored for their local environment, you start to see that those environments are way more innovative.
Read the full transcript
17:07The Life Scientific:And I learned so much by working with my colleagues who are using these things in health or agricultural monitoring. you're not looking at the whole system end to end from scratch. What they do is they go from the farmer's field to the Ministry of Agriculture or from the health centre to the Ministry of Health and put in a new solution end to end, taking into account the needs of all those different people. Now, it's extremely inspiring, but it's incredibly intimidating because you realise the diversity of those needs.
17:34Neil Lawrence:And you talked about a big issue here. It's not that people can't think up and create the technology, say, developing a new app. It's about whether they can actually manage that app and keep it running long term.
17:46The Life Scientific:That turns out to be the big challenge. There's great innovative individuals. But to build and sustain these applications in practice requires so many software engineers, so much control to maintain and explain the software in practice. And that's beyond the local capabilities to produce that number of people, because the way we build and deploy software is so poor quality.
18:09Neil Lawrence:So the biggest obstacle is actually about how the technology is managed and the fact that it tends to be dominated by the big tech players and the billionaires.
18:18The Life Scientific:Yeah. So from the big tech companies perspective, this isn't a challenge. So they're very little incentivized to directly deal with this. And in fact, what they are incentivized to do is enter new markets quickly, building technologies that perhaps are not as robust as we'd like them to be because they want to be the first mover and dominator market. They care less about the sustainability of those tools. And that's where the dystopia is coming from. We really need the people who understand the problem to be the ones that are controlling and building those tools, like my colleagues in Africa. And a lot of my research now focuses on, well, what are the barriers to that?
18:54The Life Scientific:Because if I can create a world where two African master's students can do their own ride sharing app like an Uber for Kampala, then I know I can also create a world where a couple of nurses in the UK can also change the way that they're doing data entry to ensure that they can spend more time with patients. I believe you refer to the problem as the digital oligarchy. The digital oligarchy, precisely. And I was referring to it like this 10 years ago, and it's depressing to what extent that's come through, that as you increasingly own this new digital information infrastructure, you have an increasing amount of power.
19:30The Life Scientific:It's like we've gone back to the scribes of Mesopotamia. When people first invented writing, the scribes had an enormous amount of power because they were the only ones who could write things down and read what was written. today that power has shifted into the digital ecosystem the software engineers are the modern scribes their guilds are tech companies that's the dystopia that's with us today and that's the
19:52Neil Lawrence:thing we need to change well let's get back to your story neil in 2012 up pops this amazing opportunity for a car lover like you you're asked to be a consultant for the ferrari f1 team
20:07The Life Scientific:Yeah, that was amazing. I mean, how did that come about? It came about through Connections, an old colleague of mine at Microsoft. And it was shortly after Ferrari had conceded a world championship through poor strategy. So they were rebuilding their strategy team. And I was asked to come over and give a bit of support. So what sort of thing you do?
20:29Neil Lawrence:Because I know Formula One these days is all about strategies, all about this huge amount of data that's being processed, everything from tyres to use, when to go pit stops, there's so much that's not actually involving the drivers themselves.
20:42The Life Scientific:Well, so much of it was just to do with the power of Gaussian processes to handle uncertainty, because you want to be making decisions that are considering all the possibilities. And it's not good enough just to look at one way, the most probable way, because there's so much uncertainty about rainfall or cars crashing, you need to take into account everything for sort of split-second decision-making. and watching the race it's all about the excitement and well i always think of murray walker but you know the commentator's voice rising with the excitement but in the pit garage even in the italian uh pit garages the interface is there to sort of try and show you exactly what's going on and remove some of the excitement and in some sense that maybe removed some of the love of the sport because you just see well it's actually just some cars going around in a circle but it was amazing to get to see Nigel Mansell's Ferrari and Alan Prost's Ferrari and all these sort of things.
21:34Neil Lawrence:Well, meanwhile, you were about to make a big career move. In 2016, Amazon bought a startup that you'd co-founded, which is based on your research into AI and this idea of uncertainty around what the models know and don't know. And you left the University of Sheffield to join Amazon as Director of Machine Learning at their office in Cambridge. Was that a tough decision?
21:55The Life Scientific:very tough i really loved and still love sheffield we're still sheffield united season ticket holders and very difficult for my kids but it was just the best way to try and understand what was going on with managing these large-scale deployments because amazon have to manage a supply chain there's information moving between the real world and their computer systems a bit like the ferrari f1 team but going on constantly every day hundreds of millions of pounds being spent and i knew if i could see how they'd cracked it, I would be able to take those lessons and deploy them in the areas that I was so passionate about in terms of particularly the medical field.
22:33Neil Lawrence:I hear when you moved to Cambridge, your wife gently suggested it was time to get rid of a prize collection you'd had for years of various old computers.
22:43The Life Scientific:Yeah, that was a big wrench too. But I suppose we all had to have a wrench. So I'd collected them all during my PhD, buying these old, I just, because I grew up with them. It's a nostalgia thing. But they were all donated to the Centre for Computing History in Cambridge. So I can still go and visit them occasionally.
23:00Neil Lawrence:So life at Amazon, you were working on developing machine learning solutions for them. But how different was it applying this tech in the commercial world?
23:08The Life Scientific:Very different, very interesting, but very different across different parts of the company. So some of what we were doing was in AWS. AWS being Amazon Web Services. Yes. Some of it was supporting them in understanding how drones could take off vertically and move to horizontal. Other of it changing the voice of Alexa, giving emotion to the voice of Alexa. But the thing that I was most interested in was the supply chain, because that was the big complex system where the real world keeps punching you in the mouth, as it were. So whatever you imagine things are going to happen, this is where you sort of see real time the type of problems that were happening in the post office.
23:46The Life Scientific:But if we didn't pick those problems up, then it was costing the company a lot of money. And the company was instrumented to spot that.
23:54Neil Lawrence:Earlier, you spoke about the disproportionate and unhealthy control of data by the digital oligarchy of big tech.
24:01The Life Scientific:Isn't Amazon part of that? Well, absolutely. And I think one of the key things is to sort of understand that there's a lot of incentives driving that. This is systemic, right? So if you take these big tech players away, something else will just fill that zone unless you've got on top of why this is happening. So there's two things that were interesting there. One is to see how you manage these systems at scale. So the only way you can do that is to go to these companies and take those ideas out. But the other thing is to get a better understanding of why. Why is it that these large companies tend to dominate?
24:32The Life Scientific:And a lot of it, as you say, is about control of data. And I'd written about the digital oligarchy before I joined Amazon. And then I was able to see it operating in practice. But it's not people sort of being evil and rubbing their hands together. They're trying to do a good job, or mostly they're trying to do a good job. Maybe some of them are being evil. But understanding the systemics of that and how we can ensure that small, medium, large UK companies are part of that because it's not just small businesses and hospitals that are being excluded. It's our largest FTSE 100 companies that are being excluded by the digital oligarchy.
25:06Neil Lawrence:Well, overall, it sounds like you had a positive experience there. Nevertheless, a few years later, you made the move back to academia, to Cambridge University, where your dream job came up, the inaugural deep mind professor of machine learning. And I think a big part of your work now is about deploying machine learning in all sorts of practical ways across society.
25:29The Life Scientific:Yeah. So much of what I'm interested in now is exactly that question of what does it look like for our local councils or a local hospital, or I work with a group called AI in education, to deploy this technology in a way that is supporting the teacher, the nurse, the council planning official, rather than the sort of one-size-fits-all way that we've had technology thrust upon us so far.
25:53Neil Lawrence:Of course, all this new machine learning technology does rely on the power of data, ultimately. And you've long been an advocate for data transparency and privacy. How worried should we be then about how our data is used?
26:07The Life Scientific:I think it's very worrying, and it's hard to bring home to people because it's not a sort of vivid worry. It's a sort of creeping worry. And, you know, we go all the way back to Cambridge Analytica when people suddenly became very aware of it. I initially thought, oh, it's happened now. So it's a bit late, but at least people are aware. But we keep making the same set of mistakes again and again. And it's very disturbing that the lessons of the past aren't learned and integrated. And a lot of it is because, of course, there's a desire to get the best out of these capabilities and rush them into deployment.
26:38The Life Scientific:But unless we get the data foundations right, we'll keep getting these scandals.
26:42Neil Lawrence:And to that end, you've co-founded the Data Trusts Initiative.
26:46The Life Scientific:Yeah. So the idea is that it's all very well talking about how data should be used, but how to use data in practice, you need to be at the coalface. So we actually gained some funding and distributed it to pilot data trusts so that we can learn from those tests, because otherwise the whole conversation just becomes an academic one.
27:05Neil Lawrence:Your book, The Atomic Human, came out last year. And in it, we've sort of mentioned this before, you play down concerns that machines could ever really take over from humanity. But there's still a lot of confusion and even mistrust around AI. What would you say to the average person about how this technology will change their lives?
27:26The Life Scientific:So to just try and imagine our future is extremely hard. But there's a few things I think we can say that the precious thing in the future will always be human attention. That's the limited thing in the economy. That's the limited thing that we can only share 2 ,000 bits. If you're choosing to spend time with me, I'm choosing to spend time with you, your listeners are choosing to spend time listening. That's an incredible gift and that will never go away because that will always be in limited supply. Humans will also never stop being interested in other humans. We're fascinated by each other. It's part of how we communicate and how we move forward.
28:03The Life Scientific:And then the final thing is that, you know, companies will never stop trying to corner that market. So there's a responsibility on each of us to celebrate those things about us we enjoy, celebrate those interactions, use these tools to the best of our ability and try and bring about a world where we remain in control of those interactions, because that's the pleasure of being alive.
28:26Neil Lawrence:Neil Lawrence, thank you very much for sharing your life scientific. Thank you, Jim. And thank you for listening. I'm Jim Akalili and my producer is Lucy Taylor. Lives Less Ordinary is bringing you a special series of extraordinary love stories from all around the world.
28:43The Life Scientific:In India, love at first sight feels more psychotic than arranged marriage.
28:46Neil Lawrence:Love stories that break all the rules. We came into our relationship with a lot of complicated pieces. Stories that show that sometimes love is all you need. Lives Less Ordinary love stories from the BBC World Service. Listen now. Search for Lives Less Ordinary wherever you get your BBC podcasts.
From the publisher
When you think of Artificial Intelligence, does it inspire confidence, or concern?
Although it's now generally accepted that this technology will play a major role in our future, a lot of conversations around AI and machine learning come back to the argument over us losing control and robots taking over.
Happily, Neil Lawrence has a more optimistic view of the power of AI, and how we might navigate the potential pitfalls. Neil is the DeepMind Professor of Machine Learning at the University of Cambridge, and over the course of his career has been involved in deploying AI and machine learning in both academic and commercial scenarios, with a stint at Amazon as well as working across fields as varied as movie animation, Formula 1 strategy, and medical research.
Speaking with Professor Jim Al-Khalili, Neil says ultimately his efforts are all about making a difference to our everyday lives - and that we need to learn how to embrace AI, albeit with a healthy dollop of scepticism; not least when it comes to how our data is used, and the power of 'the digital oligarchy'...
Presented by Jim Al-Khalili Produced for BBC Studios by Lucy Taylor Reversion for World Service by Minnie Harrop