The prospect of Artificial General Intelligence (AGI) now being considered capable of performing intellectual tasks that previously could only be performed by humans, has raised both expectations and concerns. Interestingly, one of the most nuanced analyzes of the pros and cons of AGI has been done by OpenAI CEO Sam Altman himself. In one of his blog posts, Altman has expressed such skepticism, which is rarely seen among people who are optimistic about technology. He wrote that ‘We want AGI to give humanity the power to flourish maximally in the universe.’ But he also acknowledged that doing so would involve ‘dealing with huge risks’. Since the Enlightenment period, instead of superstition and conservatism, the skepticism of science has increased in humanity. There is also an increased belief that scientific progress is inherently beneficial. This thinking says that even though discoveries like the atomic bomb may be destructive, science, which gives a deep understanding of nature, is undoubtedly capable of improving human welfare. However, what is often ignored is that unwavering faith in science can itself become a conservative belief. In a 2024 paper, Jorgen Weibull and I studied this trend using game theory. When humans communicate with each other under such uncertain conditions – in which what is rational for one person depends on what is rational for another – a new scientific discovery can lead to a situation similar to the ‘Prisoner’s Dilemma’ theory. In this, each individual’s personal, rational behavior ultimately leads to outcomes that are worse for everyone. The rise of AGI may prove to be a similar phenomenon. And that’s why Altman’s skepticism is justified. They say that AGI is capable of giving incredible capabilities to every person, but it can also cause serious accidents and social disruption. In such a situation, the challenge is to minimize the risk of ‘knowledge curse’. Since the future is uncertain, the best we can do is make educated guesses and adopt prudent prevention measures. These challenges cannot be left to the private sector alone. To deal with these, some degree of redistribution of resources and international cooperation will be necessary. Perhaps the most serious consequence of AGI will be a decline in labor demand. But this does not mean that people will have nothing left to do. In fact, they can enjoy leisure more than ever before. But if labor is defined in bookish language as work done in exchange for remuneration, then the world of AGI can make labor irrelevant, in which machines perform better than humans in almost every field. Since much of the adult population depends on labor income, such a change could deprive billions of people of their livelihoods. That’s why Altman has advocated Universal Basic Income (UBI), which would guarantee a minimum standard of living to every person. But beyond the impact on jobs and incomes, AGI could also open the way to global authoritarianism. If control of these systems were concentrated in a few hands, a few billionaires could gain unprecedented power over all of humanity. Having universal access to AGI is only part of the solution. Equally important is to ensure that there is no excessive centralization of money and power. If control of these systems were concentrated in a few hands, a few billionaires could gain unprecedented power over all of humanity. It is also important to ensure that there is not excessive centralization of money and power. (These are the author’s own views)
Source link
Kaushik Basu’s Column: Who should control ‘super smart’ systems?
