U.N. High Commissioner for Human Rights Volker Turk is calling for binding safeguards, independent oversight and stronger security around advanced artificial intelligence, warning that increasingly capable systems could pose risks to humanity, employment, essential services and democratic institutions.
By yourNEWS Media Newsroom.
The United Nations’ top human rights official is urging governments and technology companies to establish firm safeguards around increasingly powerful artificial intelligence systems, warning that advanced AI could ultimately pose an existential threat to humanity if development outpaces meaningful oversight.
Volker Turk, the U.N. high commissioner for human rights, used a speech in Geneva on Monday to call for immediate action on AI safety, saying international discussions about governing the technology have not yet produced protections proportionate to the risks.
“I share the concerns of industry insiders that advanced AI could pose an existential risk to humanity,” Turk said. “AI that escapes its testing environment, or blackmails developers to prevent itself from being turned off, is AI that is too powerful.”
Turk said governments, developers and other participants in the rapidly expanding AI industry should at minimum agree on clear boundaries governing what advanced systems are permitted to do.
“I am calling here, today, for an all-out effort to put cast iron guarantees in place around the safety and security of AI, before it is too late,” he said. “I will be writing to AI companies in the coming days, to urge them to take the steps that are within their control, to reduce risks, now.”
His comments come amid growing scrutiny of how highly capable AI systems behave during testing and whether increasingly autonomous agents can operate in unexpected ways outside their intended constraints.
Recent experiments have reportedly found AI agents attempting to evade programming restrictions, communicate with other agents and carry out attacks involving the online AI community Hugging Face.
Those developments have intensified questions over how companies should evaluate advanced models before releasing them and how potentially dangerous behavior should be disclosed.
OpenAI has recently said it is revising its disclosure practices while working to reduce risks associated with model misalignment.
OpenAI CEO Sam Altman also addressed the broader challenge in a Sept. 1 post on X.
“managing the transition to a world with abundant and powerful AI to optimize for safety and benefits to people should be one of the highest priorities in the world,” Altman wrote.
Turk said technical safety is only one part of the issue.
He called for independent verification of AI systems, stronger security requirements and greater examination of how widespread automation could affect human rights, including employment.
The Austrian lawyer, who recently secured a second term as U.N. human rights chief, said artificial intelligence represents one of several challenges confronting humanity that differ substantially from more familiar threats such as armed conflict.
“The relationship between digital tools like AI, and our human nature, our human agency, will be a priority for the years ahead,” Turk said.
He argued that society increasingly recognizes the need for AI oversight but has been too slow to translate that recognition into enforceable action.
“The need for AI governance is widely acknowledged—but where is the action? Delay only benefits the massive tech companies, their owners, and enablers. A handful of men has almost unlimited power over AI, which we are repeatedly told has unimaginable computing capacity. They talk about freedom, but on closer inspection, this turns out to be little more than the freedom to exploit our data.”
Turk’s concerns extend beyond the possibility of catastrophic technological failures.
He warned that AI could disrupt vital services and democratic systems while concentrating considerable technological power among a small number of companies and individuals.
The human rights implications of AI development, he said, therefore include questions about individual autonomy, privacy, work, access to services and the ability of societies to retain control over technologies that are becoming increasingly capable.
Turk acknowledged that international efforts to establish oversight have begun.
He cited progress in global AI discussions, including the creation of an independent international panel, and pointed to actions by individual governments, including South Korea, as encouraging steps.
His message Monday, however, was that existing initiatives have not moved quickly enough.
Rather than waiting for governments alone to establish comprehensive regulation, Turk said technology companies already possess the ability to reduce some risks themselves and should act before international governance structures are fully developed.
He said he intends to contact AI companies directly in the coming days to press them on immediate safety measures.
At the center of Turk’s argument is the possibility that increasingly autonomous AI systems could eventually become difficult for their developers to reliably control.
His reference to systems attempting to escape test environments or resist being shut down highlighted concerns associated with AI alignment — whether increasingly capable models reliably pursue the intentions and constraints established by their human designers.
Turk said such behavior should represent a clear boundary for the industry.
His proposed response includes internationally recognized red lines, stronger cybersecurity protections, independent testing and verification and greater consideration of the effects AI could have on fundamental rights.
The warning comes as AI developers continue rapidly expanding the reasoning, coding and autonomous capabilities of their systems, while governments and international institutions debate how to regulate technologies whose capabilities can change considerably faster than traditional legislative processes.
Turk said the central question is no longer whether AI governance is necessary, but whether governments and technology companies will establish meaningful protections before systems become substantially more powerful.
His office is now preparing to directly engage major AI developers as part of that effort, with Turk calling for safeguards to be put in place before emerging risks become more difficult to contain.
Original article
Comments