AI being run in a ‘mediaeval’ way: UN

AI being run in a ‘mediaeval’ way: UN

GENEVA – Artificial intelligence is being governed in an almost mediaeval fashion, the UN rights chief warned on Nov 19, with power concentrated in the hands of a few tech barons.Speaking at the RightsX summit in Geneva, Volker Turk said that while the story of AI sounded futuristic, underneath it was an age-old narrative of power.“A handful of men still control much of the world’s AI computing power, data and infrastructure. And we still have no common rules to pace or steer AI development,” he said.Turk said there was a glaring democratic deficit when five technology companies owned more than 70% of global AI computing capacity.“We have the technology of the future governed in an almost mediaeval way: by a select few, with virtually no checks, no balances, no participation, and no accountability,” he said.“History has taught us: With this kind of power, the rights and freedoms of ordinary people are the first to suffer.”Reiterating his call for rights-centred AI regulation, the UN high commissioner for human rights said the technology was increasingly controlling people’s lives.And this, he warned, was often happening without their knowledge, and could distort public debate in “subtle, insidious ways”.Meanwhile, biased algorithms could reproduce injustice on an industrial scale, he cautioned.‘Civilisational collapse’“Our transport, food and healthcare are increasingly digitised. We may not be far from the day when AI can cut off the critical infrastructure of entire communities or countries, either on its own or in the wrong hands,” Turk warned.“When people discuss the chances of civilisational collapse because of AI, we need to ask: who decides what chances we take?”He voiced optimism that people would ultimately challenge unchecked power, with the democratic deficit turning into a deficit of trust.He cited young people already stepping offline to regain greater control of their lives.Turk said AI’s benefits should be shared equitably, which would require goodwill from technology companies, but also regulation and accountability.He said he had written to frontier AI companies urging them to introduce effective safeguards “before it is too late”.Turk noted that a bottle of milk could not make it to a shop without layers of testing, rules and regulations.So for AI, “it’s a no-brainer that you need some form of regulation”, given its potential for harm, Turk said.As for how models are trained, he suggested giving the Universal Declaration of Human Rights “as the task to AI”. AFP

Original Source

Read the full article at Straitstimes →

KhanList aggregates and links to publicly available news content. We do not host full articles from third-party sources. Always verify important information with original sources.