At the RightsX summit in Geneva, UN High Commissioner for Human Rights Volker Turk sharply criticized how artificial intelligence (AI) is governed today. According to him, the situation resembles the Middle Ages: power is concentrated in the hands of a few technology barons, and ordinary people are deprived of the ability to influence what happens.

Turk noted that just five companies own more than 70% of the world's computing power used for AI. This creates, in his words, a "glaring democratic deficit" - that is, a lack of public participation in decision-making. "The technology of the future is being governed almost in a medieval way: a narrow circle of the chosen, with practically no checks, balances, participation or accountability," he said.

According to the High Commissioner, history teaches that when power is so concentrated, the rights and freedoms of ordinary people suffer first. He warned that AI is increasingly influencing people's lives, often without their knowledge, and can distort public debate "in subtle, imperceptible ways." In addition, biased algorithms (that is, programs that can reproduce injustice because of errors in data or rules) are capable of replicating injustice on an industrial scale.

Turk also pointed to growing risks for infrastructure. Transport, food supply and healthcare are increasingly digitized. "We may not be far from the day when AI could shut down the critical infrastructure of entire communities or countries - on its own or in the wrong hands," he warned. In this connection, he asked: "When people discuss the probability of civilizational collapse because of AI, one must ask: who decides what risk we are taking?"

Despite the alarming forecasts, Turk expressed optimism: he believes that people will ultimately challenge unchecked power, and the democratic deficit will turn into a deficit of trust. He cited the example of young people who are already refusing to be constantly online in order to regain control over their lives.

The High Commissioner stressed that the benefits of AI must be distributed fairly. This requires the goodwill of technology companies, but also regulation and accountability. Turk said he had sent letters to leading AI companies urging them to introduce effective safeguards "before it is too late." He compared the situation to a bottle of milk that does not reach a store without multi-stage checks and rules. "For AI, it is self-evident that some form of regulation is needed," he concluded, referring to the technology's potential harm. As for training models, Turk proposed setting AI the "task" of following the Universal Declaration of Human Rights.

Source: Tech Xplore