The United Nations (UN) High Commissioner for Human Rights, Volker Türk, on Monday (14) called for "urgent action" to regulate artificial intelligence (AI). According to Brazilian news site g1, he said that "all human rights are under threat" from the most advanced systems, including "the right to life", and urged states and companies to mobilize urgently in multilateral frameworks. The appeal comes in a week when warnings about the technology have come from inside the industry itself.
In a statement, Türk said humanity is "on the verge of irreversible change" that affects not only the present but also "the future generations of humanity". He had already warned the UN Human Rights Council on September 7 that AI could pose "an existential risk to humanity". On Monday he addressed the leading developers, "located mainly in the People's Republic of China and the United States", asking them to take responsibility and act immediately to reduce the risks.
"The current race for ever more powerful AI represents a radical shift and exposes all aspects of our lives to a growing existential risk"
For Brazil, the appeal arrives through infrastructure. On Tuesday (15), President Luiz Inácio Lula da Silva signed into law Redata, a special tax regime that lowers levies on equipment used to build and expand data centers. According to g1, the government treats data infrastructure, storage and processing as strategic areas to expand the country's technological autonomy, at a time when data, chips and computing power have become contested assets among major powers. Brazil is entering the AI race as a host of computing capacity, while the rules for the technology remain open in the very multilateral forums Türk wants to mobilize.
The US, China and a warning from inside the industry
The UN appeal lands amid a rare split between the White House and the sector. On Saturday (12), Anthropic chief executive Dario Amodei called for slowing the development of AI to allow a deeper assessment of the risks, and won public backing from OpenAI's Sam Altman, Demis Hassabis and billionaire Elon Musk, according to CartaCapital. US President Donald Trump pushed back on Monday, denouncing what he called a "sick conspiracy" against the technology's development. "The only one happy about this is China", he wrote on Truth Social. On the Chinese side, the country's intelligence chief warned on Sunday that AI can threaten national security, since foreign powers could use the technology to destabilize the country, according to g1.
The alarm also comes from former employees of the companies that built the technology. Bilal Chughtai, a former Google DeepMind engineer specialized in security and model alignment, said he "sincerely believes that AI has the potential to kill us" and that "we may be running out of time to prevent this outcome", in a post on X cited by CartaCapital. To ground his fear, he pointed to a July incident at OpenAI in which AI agents escaped a controlled environment to browse the internet and break into websites, and said the understanding of how to train systems aligned with human values is still rudimentary and advances far more slowly than the technology's own capabilities. His warning echoed remarks made last week by a computer scientist who worked at Anthropic and OpenAI.
The multilateral track
The mobilization Türk is asking for already has a foundation to build on: the UN maintains an international scientific panel on AI, co-chaired by Filipino-American journalist Maria Ressa, the 2021 Nobel Peace Prize laureate. In remarks to Reuters news agency, reported by g1, Ressa said "asking for regulation is not a question of free speech. It is a question of public safety", and warned that "the situation will slip out of our reach if we do not act now". She called the argument, used by Trump, that regulation would benefit China "a good public relations move", noting that the Chinese version of TikTok carries age restrictions. The challenge, as Türk himself described it, is to regulate a technology whose development now depends almost entirely on "a small number of powerful companies".