{"text":[[{"start":5.65,"text":"Amid the many debates over AI, everyone can agree on one thing: this marvel of a technology throws up great and unprecedented uncertainties. The possible endgames put forward by serious observers range from a productivity utopia to the annihilation of humanity."}],[{"start":21.75,"text":"That uncertainty need not be paralysing. Instead we can take it as a source of focus: it means that how we address the AI revolution should be seen as a question of how we choose to manage uncertainties, and what tools and principles we adopt to harness risk for the best result."}],[{"start":39.9,"text":"Recently even Donald Trump’s White House — hardly an AI-sceptic administration — has shifted towards putting at least some regulatory constraints on the most powerful models before they are let out in the wild. In contrast, Argentina is trumpeting a different approach. Its president, Javier Milei, declared in the FT this week that his country would commit not to impose regulations on what AI algorithms will be permitted to do, and to introduce a legal category of “non-human corporation” — companies run entirely by AI."}],[{"start":72.85,"text":"The move may be a gamble to attract AI developers to a country in search of a more stable growth model. As Milei points out, the history of the limited liability company illustrates that encouraging capitalists to do risky things can bring about good outcomes. Peter Thiel, the billionaire tech investor, seems taken with Milei’s approach; he has reportedly just moved his family to Buenos Aires. "}],[{"start":96.6,"text":"A free-for-all for AI will be too much to stomach for most jurisdictions, however. In what can only be called the opposite corner from Thiel, given the investor’s fascination with the antichrist, Pope Leo offers an explanation why. His papal encyclical on AI is no Luddite tract; Leo welcomes the potential of AI and other technologies to “serve integral human development and the care of our common home”. He insists, however, that risks pertain not only to the outcomes that technologies pursue, but the vision from which they do so, and what we humans become in the process — in other words not only the “what” but the “why” and “how”."}],[{"start":138.7,"text":"Leo reminds us while AI may surpass human intelligence, they are not the same thing. AIs “do not know from within what love, work, friendship or responsibility mean”. Machine learning “does not imply inner growth”. He warns in particular against so-called post- or transhumanist views, because these attempts to improve humanity see human limitations as flaws to get rid of. "}],[{"start":160.95,"text":"The pope’s counterpoint that humanity flourishes “not despite limitations but often through them” is one that many people are becoming more aware of in the case of “cognitive surrender”: the realisation that making things easier through AI can diminish rather than enhance our abilities. It is in a similar vein that the FT commits to always keeping human judgment at the centre of our journalism."}],[{"start":185.5,"text":"In some ways, these contrasting approaches — one chafing against limits, the other identifying humanity with them — can perhaps be reconciled. But that requires accountability to be located somewhere. From papal reflections to the sci-fi works so beloved by tech elites, the founts of human wisdom agree that humanity will not long tolerate a system producing injury without redress."}],[{"start":210.2,"text":"So one can envisage permissive legal forms if accountability is assumed by morally mature individuals, companies and governments. But then it is incumbent on us as individuals and collective cultures to retain our own judgment of how our society fares, and avoid the path of least resistance that leaves uncertainty to be managed by the machines."}],[{"start":236.75,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1780806486_4276.mp3"}