daron acemoglu reckons technological progress often reinforces inequality and exploitation. so, how do we ensure the tech revolution works for the many, not the few? daron acemoglu, in massachusetts, welcome to hardtalk. thank you, stephen it's a delight to be on the programme. thank you. a great pleasure to have you here. you have just written a book, power and progress, which goes through the impacts of some

of the great technological shifts in human history over the past millennium. would you say that artificial intelligence, the supremacy of the algorithm, represents one of those transformational moments for humankind? i'm not sure that it does. but there is a fair chance that it will have transformative effects. it is all commonplace to compare it to the fire, or to the rise of industrial machinery starting in the 18th century, it remains to be seen whether it will be as important for our productivity and our lives, but we are already seeing some of its deep effects on how we organise our democracy, and inequality in modern nations. maybe you are not quite so sure about how dramatic its transformational impact will be, but you certainly seem to be pretty sure that you're sceptical. you say ai will fuel inequality, disempower workers,

and, choke democracy? how? ai is a continuation of a trend that started perhaps around two years ago, where we have been using digital technologies, for changing how production is organised, who controls information and how we communicate. both digital technologies in general and ai have great promise, we can talk about some of the specific ways in which ai could be transformational for knowledge work, for example, but in practice, they have actually boosted inequality because they have been used for automation increasing surveillance and also centralising information. i guess all that is true but it depends on how you look at it whether you are a glass half full or glass half empty guy. for example, not so long ago, i was talking to a former president of estonia, and he was telling me with great pride about the degree to which his

country now is entirely online, that all citizens are digital citizens who do everything from political interaction to interaction with their healthcare providers and their doctors, online. he said, you know what, we believe that this enhances our democracy. absolutely, and i think estonia has done a good job and that is where the difficulty is. when you think about the current effects of technology and where the future will take us, you have to hold two conflicting thoughts in your mind at the same time. one is that we have tremendously benefited from industrial technologies. we are healthier, more comfortable, much more prosperous than people used to be 300 years ago and there is tremendous progress that could be beneficial for humans. but, conflicting with that, there are also a lot of collateral damages, that have been created on the path

of so—called progress. and there are many dangerous directions in which these technologies are being pushed. in that direction, you advise us to think carefully about history. and you talk about, for example, the late 18th and 19th century industrial revolutions in the united kingdom and western europe and you say in the first phase of that industrial revolution, many workers suffered stagnant incomes and deterioration in their living standards and conditions, and extreme exploitation. and you could point to the fact that, for a while, children were employed in the mills in the industrial mills and you could point to the fact air pollution was a terrible problem. this is all true, but does any of that have relevance to the transformational impacts that might come with this digital ai—driven evolution?

there is no doubt that al is very different from the early phase of the british industrial revolution but there is a lot we can learn from the past because the same sort of choices and the same sort of debates were prevalent, and, also, there are some systematic ways in which technology gets captured by current elites to be used in their interests, and against the interests of many people. so what was so important about the british industrial revolution was that it was truly different from what came before it. it was driven by the vision of a middling sort of people who were rising from relatively modest background, and trying to transform society, and that sounds really exciting. but at the end, a lot of it, exactly like you said, was pretty bad for the working classes, for about 80 years, 90 years. the early phase is characterised by steam engines but what that meant was children as young as five

or six were sent into horrible minds to work 18 hours a day, seven days a week. life expectancy fell sharply because cities where people started gathering, were so unhealthy, so polluted, infectious diseases run rampant and real wages did not increase. the factory system was a pretty draconian place, with very harsh discipline. long hours, little economy. definitely, but as you say all that and i can almost feel and see and hear the giants of tech in the united states, from bill gates to elon musk, rolling their eyes and saying, "come on, "get real, "professor, what we are delivering to the people right "now are technologies that are actually in many ways, "opening up knowledge and information, "in extremely egalitarian ways to humanity." is that not the truth?

well, i don't think that is the truth. that potential was there, there were many inspired people in the 19605 and 19705 who thought that computers, personal computers, digital technology more generally would be liberating for workers and citizens but that is not the way things have turned out. most of the digital technologies are controlled by big corporations, inequality has skyrocketed in many countries, the uk and the united states included. in the united states when you look at real wages, they have actually declined for people with less than college, for over a0 years. are you not guilty of scapegoating technology, if you blame technology for that? you can blame all sorts of other things, including a whole free—market capitalist ideology, that perhaps one could identify with reagan and thatcher and a neo—liberal

movement in the 1980s, that has got nothing to do with technology. that's ideology, it's not technology. in the book with simon we do put a lot of emphasis on the institutional changes and the changes in division of corporations, the shareholder value revolution, which said it is ok to just squeeze the workers harder so we can give a better return to the shareholders. but, actually, my own research shows, that the application of digital technologies was central to how the wage structure changed in the united states. you cannot understand what happened in the united states without robots, and numerically controlled machinery and office software displacing workers, especially middle skilled workers from their previous tasks and jobs. let's bring this down to level we can all understand, our own lives and how we operate. i can imagine there are people watching and listening to us who are now using chatgpt to help them do a work presentation

or if they are a student, help them to write the next essay and they are embracing that technology they are finding it extraordinarily useful for them. now, you appear to be saying, you know what, be very careful because this is actually doing structural damage, notjust to your life but to the society and economy in which you live? i think the answer to your question comes back to the point i made earlier. ai has great promise and potential for great damage. the generative al or the large language models could be a new tool for augmenting what we do in knowledge work, in white—collar work, journalism, academia, it's not currently going in that direction because, actually, gpt�*i or chatgpt have very impressive—looking results, but deep down, they are not super useful for knowledge workers at the moment. but at the same time, many of these

things are already being used for centralising information further, and replacing workers that used to do simple analytical tasks such as writing tasks or summarising news and so on, that is only going to expand. it is the choice we have to make, how we will develop these technologies, and how we will use them, that is the critical one, especially when we are at the cusp of more and more investment in this generative ai. but as an economist you know that these technologies, including ai, are already being employed across the industrialised world, and if you also know that if you look at united kingdom or the united states, have almost structurally full employment. yes, we have people out of work but not very many and job markets are very tight. and if you look at, for example, consultancy groups like pwc, they've analysed what they think will happen as a result of ai, they wrote a report called, will robots really steal ourjobs?

they said some jobs will go as a result of automation by ai but they will be outweighed by the creation of newjobs that will come, thanks to the expansion of the economy as a result in large part of the uses of ai. so, net result? they say, expansion ofjobs, not loss ofjobs. yes, i wish that were true. the potential is that the gain there, when you look at earlier phases of industrial growth, the three decades that followed world war ii in the uk and europe and united states, that's exactly what happened. there was rapid automation but at the same time, new tasks were created for workers that notjust generated employment but actually ensured that inequality fell, real wages rose very rapidly. but that was a choice about institutions, just as you pointed out, do we completely give in to the wishes of

the ceo and shareholders, and it was also a very important choice about the direction of technology. the current direction of ai will not create a jobless future in the next three decades, we will not be all sitting at home and playing video games, but if it goes along the same path of the previous digital technology, it will boost inequality quite substantially. you seem to be proposing we need to learn from the past and you weave history and economics and politics altogether, you say learn from the past and look at how we have ameliorated in the 19th century the worst impacts of industrialisation by workers forming unions. political parties representing the workers, rules and regulations been bought in and you say we need to do the same again, in the 20 century? we need to learn how wisely to regulate and engage workers in the process of, responsibly

deploying ai. so how, when you look around the world, how do how you actually see that working, in practice? well, i don't see it working too well at the moment. i think the key is, you need a 2—pronged approach. in the past that has happened that has happened without a map, perhaps we were lucky at times, but we need both to build institutions for better voice and better sharing of gains, but we also need to push technology in a more pro—human direction. i think both of these are very difficult things and we have been failing at them. when i say about technology being in a more pro—human direction, that means use technology to increase workers�* contribution to production, create new tasks for them, provide better information for them, not just automation, notjust monitoring. i also mean using it for citizens, information, for example, in estonia. when citizens can participate better in political decision—making, we know from estonia and taiwan, it works quite well. but look at what facebook has done for social

media and other social media outlets. what generative ai is likely to do with deep fakes and more powerful disinformation is a very different path. and the institutions aspect, you put yourfinger on it, completely rightly, it is critical but we are in the middle of one of the strongest slides of democratic strength all around the world. how will we build new institutions, especially when the labour movement of the future has to be very different from the labour movement of the past. all of those are key questions that we have to confront. yeah, but famously, companies like google weren't too keen on unions, and one could argue that people like elon musk are fundamentally individualists and maybe even libertarians. but the argument would run that that's why the united states has been the absolute cockpit of dynamic creativity, when it comes to tech and ai, and rolling out the very best creative new technologies, that's why the us has a competitive advantage over countries like china,

because it is much less feted, it is much less interventionist and regulated. would you not accept there's some truth in that? well, again, we have to hold two conflicting thoughts in our minds at the same time. yes, you're100% right — i don't think silicon valley, in its current form, would exist without some amount of individualism and entrepreneurial risk—taking and some of that break — move fast and break things attitude. and you're absolutely right — silicon valley is a great innovation machine and we are very fortunate to have benefited from some of the innovations that it has created. but also at the same time, that same culture of silicon valley has been the thing that has pushed us towards more and more automation, less and less attention to what the average worker needs — a sort of an elitist attitude towards technology — and i think with the advent of ai, all of that could accelerate and create a bigger disaster, so the question

is can we have the good of that energy, entrepreneurial risk—taking, without some of the bad? and the answer is yes. again, we can learn from history. united states has always been an entrepreneurial nation but technology and the corporate sector was under better institutional controls and regulation. it was no less innovative in the age of early computers, aerospace innovations, antibiotics, but it was — that energy was being channelled towards things that were better for humanity. i think we can recreate that. are you somebody who, like elon musk — actually, not so long ago, just a few months ago — has now begun to worry that al might take us to a place where we are no longer masters of the machine but the machines may become masters of us? i mean, how far do you go down the doomsday track? no, i'm not worried about that. i think the current architecture of ai is unlikely to create anything like the human mind and i also worry that some of that discussion

of existential risk really takes our attention off the more mundane and daily user of ai and digital technologies, that are inequality—generating, that are centralising information, that are boosting the inequities that exist. i think a very important point that, you know, hg wells captured, in the time machine, when he said, you know, when talking of human dominion about nature, we forget that it's always entangled with human dominion over other humans. that's the danger. we are using ai in a way to subjugate people, reduce their autonomy, reduce their agency, reduce their contribution to production, and that's being done not because machines have taken over but because a small cadre of people are shaping the future of ai. what we see right now in the world economy is a fundamental shift away from assumptions

of globalisation, open trading networks, to something where borders are much more important, where national interest is focused upon and where, frankly, many governments, democratic and non—democratic, are talking in terms of protection, protecting their own interests. do you see this as an extremely worrying shift or not? do you see this as an extremely worrying shift, or not? i think we could overdo it, and a collapse in the world trading order would have bad economic consequences and would probably make international cooperation much harder. but i think our previous preconceptions about globalisation were also not completely on target. just like technology, there was this view that let globalisation rip. if china wants to export cheap products, let all of the producers in the united states or in europe go bankrupt and then, the market'll take care of it. same sort of attitude when it comes to technology — let's not worry about these inequality issues, let technology rip and then,

the market process will make sure that people are allocated to appropriate jobs and they all benefit. and what we have experienced from globalisation is that when that takes place in an uncontrolled way, it can also create a lot of hardship, so i think some degree of worrying about, you know, what globalisation is doing is fine but i really don't want the world to turn its back towards international trade, exchange of ideas. and i think the heart of the china—us tensions are better handled if we can find a cooperative environment but i think that doesn't necessarily mean that we have to completely open all borders to everything. you wrote a best—selling book in 2012, headlined why nations fail. well, we've had a decade since you wrote it, and more. have you, in a sense, reconsidered it? because your fundamental message seemed to be that, you know what? nations with strong

institutions, separation of powers, strong civil society, democracy in the end are more likely to prosper than those that don't have those pillars. well, you know, look at us and china today. maybe you got it wrong. well, of course, maybe i got it wrong, but i don't think the evidence so far disproves any of those high—level claims. inclusive institutions, which are embedded in democratic decision—making and openness to new ideas, have proven to be the best guide to innovation, and to limiting inequality, while encouraging economic growth, and i think the united states has done worse, when its democracy has weakened. same thing for europe. china is a new experiment. it has had very rapid growth under state control

and all historical examples, like china, have been afraid of innovation. what china has managed with al, perhaps, it is intensifying that, it has tried to find a path to control innovation, while also enshrining the control of the chinese communist party and chinese elites on the political system. but you're already seeing the cracks. the reason why there's so much surveillance, there's so much oppression in china, there's so much talk of nationalism is because you're not able to contain the middle class and all of these aspirations in such a centralised system, and you're also seeing that innovation is not working as well in china as its proponents claim. so, i think we're going to see a reckoning in the chinese system, but the international context, the geopolitical context you raised, that's a wildcard here. i don't know what it that's going to... crosstalk. and just a quick final thought. turkey — now, you're of turkish—armenian heritage. you made it clear during the recent turkish election

campaign that you supported the opposition, you wanted to see an end to erdogan�*s rule after he's been in power for so long and have become more authoritarian over the years. the turkish people, they had a choice. they looked at their economy, which is misfiring very badly, and yet, they decided to stick with erdogan. i think i've said that several times in the past that whenever i have a different view of what the turkish electorate�*s going to do, then erdogan, he seems to be right and i seem to be wrong. but don't forget turkey's a deeply polarised country. about 50% of the population do not like erdogan�*s regime. they were really keen on creating a new turkey that changed government and changed economic system. and erdogan has a very loyal base. part of it is because exactly the cultural issues. he liberated the more conservative, less educated, more anatolian parts of turkey. but also, he controlled state

resources and the media. turkey is still sort of a semi—democracy, but it's the weakest part of the democracy is that essentially all of tv and print media is under the government's control. and the government also controls the resources. so, for instance, the earthquake, which is really a historic event, you know, the areas where — that were most — worst hit by the earthquake actually gave 70% of their votes to erdogan. why? because he controls the patronage network. so, you have to take that broader context into account. but he's also a very skilled politician, i have to give it to him. well, daron acemoglu, there's much more we could talk about, but we have to end there. i thank you very much indeed forjoining me on hardtalk. thank you, stephen. this was fantastic. thank you very much.

hello. the warm weather is set to continue, but you might be thinking more about the rainfall by now. it's notjust our parks and gardens that are desperate for the rain, but, of course, our wildlife, and there is some rain on the horizon. if i show you what's going to be happening over the weekend, well, this low pressure here will continue to push warm air in our direction, but also a lot of moisture, and perhaps more widespread storm clouds than of late. but, again, not everybody�*s going to get the rainfall. hopefully, it will be a little more widespread, and, hopefully, it'll fall at the right time, and not on your barbecue. anyway, let's have a look at the forecast for the short term. so, a lot of clear weather, first thing in the morning. temperatures will be around the mid—teens at 7am on thursday, and then,

basically, it's blue skies, steady as she goes, through the course of the morning and into the afternoon. now, fairweather clouds will be bubbling up, and some of them will form, locally, at least, into a few storm clouds, and we could see some thunderstorms developing around the north and the west highlands. temperatures in glasgow up to 26 celsius, a warm day for northern ireland, too. also, perhaps, one ortwo showers, and the odd rumble of thunder, maybe across wales, maybe the midlands, and also in the southwest of the country, but the vast majority of us are in for another very warm and sunny day, with highs approaching the high 20s. now, here's the weather map for thursday into friday — low pressure to the west and southwest of us is nudging in. it's pushing that energy in, that moisture, those clouds, but you can see there's not too much rainfall, at all. in fact, we'll see a few maybe reaching parts of northern ireland and western areas of the uk. but, again, many of us are in for a dry and very warm day on friday, with temperatures typically in the mid—20s, and noticejust how much warmer it's also getting, because of a change in the shift in the wind direction, across the north sea coast.

now, here's the weekend. that low pressure is much closer to us, but look at that — some storm clouds coming in from the south — and this could bring some appreciable rain, to at least some parts of the country. it's not guaranteed, but i think quite a few of us will get the rainfall, so we'll have to wait till saturday, some of us sunday, and into next week. but in the very short term, it remains very warm, and, of course, very sunny. bye— bye.

live from london, this is bbc news. three days of mourning in greece after 79 migrants were killed when a boat capsized off the country's southern coast. just hours away — the long—awaited report into whether former pm borisjohn misled parliament about parties during covid is due to be released today. and more than 100,000 people evacuated out of the path of cyclone biparjoy, as it heads towards india and pakistan.