 Hello and welcome to NewsClick. Today we have with us Dr. Anupam Guha and we are going to discuss what might appear to be an esoteric subject to artificial intelligence and its impact on society. Anupam, you've been writing on this. Yes. Two issues here. One is of course artificial intelligence itself is something people don't understand. The second, of course, what are these likely implications of society? Let's start with the second issue first because I think that's causing a lot more concern that it's argued that with artificial intelligence, most of the tasks which today are regarded as quote-unquote intellectual labor, even if it is sort of peripherally intellectual, would also get automated. Irregular work could get automated which are not regular in the sense of a factory and therefore we are looking at large scale unemployment in the future. So do you think that's really the possible, that's a possible direction that artificial intelligence is traveling? I think the debate is sort of ill-framed on two different lines. First of all, we are looking at increased precarity. It's not very clear how much AI is to be blamed for that precarity and not structural issues. This is what I have been writing about for a while that AI is not really your enemy and the structure which supports the economy, which AI kind of magnifies that is what needs to be addressed over here. That's the first line I want to address. The second line is that we drive this hard line of unemployment whereas we should be talking about unemployment, precarity and precarity. So you have this new technology inspired platform economy which is forming now. So you have all these platforms which people rent out in transportation, in hospitality industries. Like people rent out their houses on certain platforms and then- Airbnb, etc. And that is also driving prices down. That has grave effects on entire economies, people are forming legislations, people are forming all sorts of things. So I think that more than unemployment, we should be talking about unemployability, this weird condition when people are forced to work two, three different gigs in what they call the gig economy. And the lines between who is employed and who is unemployed sort of become blurred because if you are talking about just a hard line of unemployment, then many economies look relatively healthy right now which they are actually not. There is a huge deal of precarity going inside the employed space which is related to technology. Economists often derisively use this phrase lump of labor. They think it's a fallacy that they think that the opposite school of thought thinks that there is only finite labor. So technology will come, labor will get over and then what will people do and then they make fun of it sort of. Well, the fallacy itself is fallacious. Even if labor is infinite, the rate of doing labor of course depends on technology and the social structures which own that technology. Let me interrupt you for a minute on that. What you're really talking about, what is employment? In fact, that's the issue. The question is, you'll pay the jobs which you keep on losing and gaining continuously. Are you unemployed or not? That's the definition of employability and being employed, precarious labor which is the condition in which case any way we have been having precarious labor in our cities for a very long time. But leaving that out, if I look at it in terms of framing it in this particular way, the amount of labor you need to produce, say, food for yourself and the basic needs of life and say a certain amount of what would be considered consumer goods or even consumption of, shall we say, leisure. All of that can be produced with the kind of technology we have and what we are likely to see in the future is going to be much lower in terms of number of people than what we would require in the past. This is one part of it. Now, if we want to resolve this contradiction in what you're raising, there are two ways of looking at it. What is employment? Let's have a look at that and pack it in that way. Why would people have to work 40 hours a week? Why can't two hours a week, four hours a week if we have that amount of technology? Why should it not go down to drive down our hours when we are looking at it? If there is enough to be produced, then do I need to fear it or do I need to fear it only because I'm quote unquote, I'm also surplus and I don't have any role in society? So is that what you're really posing? So if you have been following my writings and my talks, I am in favor of increased productivity and I'm a technological optimist. My concern is that the question of how many hours you would work, would you even work at all or be given some kind of universal basic income, which I think is a naive solution? Or would it lead to reduced work or public leisure as people are calling it? I think that's a structural question because who gets to decide the quantum of universal income or surplus value being allocated to you or how many hours you are working? Those are political questions and unfortunately this whole debate is being treated as a technological debate. Even UBI is being pushed by technologists who don't really have a solid political framework to argue this on. In the past, work hours were not a product of technological level. They were hard fought political battles. Absolutely. 40 hour week came out of huge struggles which had to be waged. So if AI does increase productivity and assures a certain amount which can be used to achieve a higher degree of sustenance, especially that is important for the case of third world countries like India. The global south needs AI. Let us not turn our faces from that. Technophobic reaction to this thing would be most unwise but who gets how much is right now decided by a very small section of the stakeholders or in fact the multi-stakeholder model is itself politically loaded and fallacious I would say. I think multi-stakeholder, multi-stakeholder model discussions are really to hide the fact that a few people have concentrated wealth in the world today. They control the governments and they also control the NGOs. So at the end of it out of four stakeholders if you will, three are on sitting on one side of the table. So this really is not a multi-stakeholder model is really not an issue. The issue that I would say that you are raising is that ultimately these questions are to be answered politically and fundamental issue then I would say that what you are proposing is that we have to look at what kind of society we want and that is not an economic or a technical answer, it is a political question. And therefore the debate of artificial intelligence versus employment really begs the question what is the future we want and as human beings what kind of society we want, would that be the way you should look at it? And what kind of society we want is not all that is often pretended to be, it is not a natural or an organic thing, basic concepts of current society, concepts like property are legal constructs, they are not natural constructs. So who owns the AI or what sort of property is AI generated production leads to or even older concepts like intellectual property etc. I think it is attempted to demonstrate that these things come from some kind of natural law whereas they do not they are social constructs. The value is socially produced even in the case of AI it is not doing some kind of magic, it is not creating stuff from nothing. The data that goes into feeding machine learning algorithms that the big data as it is called that socially generated data. It is like network of people produce data or all humans it is a script from human sources you know Facebook or social media in general is used to scrape a lot of data that is used to form opinions. Then you have these machines they scrape expert data and then you create machines which do the work of experts the data comes from human beings. The algorithms are also socially produced they are produced by scientists people like me. We study in schooling systems and university systems which are social constructs which are funded by public money a large section of it is not even privately owned. So the science of AI the technology of AI the oil of AI the data so to speak. So both the method and the content and the aesthetic form are socially generated value. But then the value gets concentrated after all that generation and I think that is problematic because everything the long chain of what goes into producing all that wealth I think that is where we should concentrate our political lines. Put it this way if I rephrase it I will say it's socially produced but privately appropriated that's the crux of it. Leaving out of the social political and economic issues which of course is the bedrock of any discussion on the future technology is only a part of the debate. Let's also look at the tool itself AI. Now a lot of the things that AI is used for it has been argued and Cathy O'Neill has this book called Weapon of Math Destruction argues that it is really sort of shall we say prejudices woven into mathematics. I would say that's probably a harsh way of putting it but it tends to replicate existing inequalities because that's what it looks like if you design an algorithm. What is likely to succeed if that's a task given to it then it is for instance take success of a application for a loan. Who are likely to succeed it is going to see all the social data and bias it in favor of existing inequalities. So would you say there is a usage of this kind in AI itself as a problem? I sort of like to oppose this line on two grounds. I don't disagree with it but I oppose it there is a difference there. First of all I would say replication of social prejudices is not something unique to artificial intelligence. Even the most primitive form of bureaucracies do that. Any bureaucracy uses data and statistics even past bureaucracies and Victorian England would do so. So if a society has structural problems they would get reflected in the bureaucracy. So the police would arrest certain demographics more than certain other demographics. AI is nowhere near new to these kind of problems. So what often people now do is that they will demonstrate that this algorithm is racist or sexist and they would say see this technology itself is problematic. I disagree. I would say the technology is reflecting society yes correct agree it. But you cannot use that to say that this technology is itself problematic. The society is problematic. Removing the technology will not improve anything. The technology is an accelerant. It will accelerate whatever you want it to accelerate the first line. The second line and this I think is my line on this is that not only is the technology not technology in general. The sole solution or culprit of structural problems I would say you cannot solve those structural problems without the acceleration provided by technology. I am not a solutionist. So I am not saying that technology is the solution to structural problems. I am saying in order to solve structural problems you need the assistance of the increased production or the acceleration the technology provides. Let me rephrase that sentence to say there is no solution without technology. That is the position I would take that while technology may be seen to be a part of the problem there is no solution which can be done without technology. I am just rephrasing what you are saying. Also politically it would be very unpragmatic because that has been tried in the past in the 70s when digitization first appeared on the world stage. What a lot of very well-intentioned people did across party lines across country lines was that they opposed it. Yeah computers are opposed in my middle class units for example. And what happened to that was that these technologies were appropriated by essentially predatorial forces. And because there was nothing to no vision of using these technologies in a better fashion. Aside from some very close circles like free software and others which came a little late. At that time the only example of using it in a more emancipatory sense was Chile's Aende where there was an attempt to create. Of course that didn't last because of military. And I am very much afraid. It's my great fear that this kind of almost naive rejection of technology is idealistic. It ignores the. It's a Luddite position and as you know. Well I would counter that. Luddites are mischaracterized. The Luddites were not actually embracing this position. It's more of a caricature of the Luddite movement. I am basically saying Luddites did break machines because they saw that rightly so as taking away their livelihood and their lives. So it would be very surprising if they did not. The question is not that. The question is not caricature and dismissing it and so on. Also recognize the problems there were. That in fact the turn against the machine there was a particular reason for it. Thanks Anupam for being with us. It's been a pleasure talking to you. It was excellent talking to you. Particularly on issues that are also close to my heart which people in news click may not realize or people watching news click may not realize. Thank you so much for watching news click. Do keep watching us. Visit our YouTube channel and our Facebook page.