(United Nations) Faced with the explosion of artificial intelligence which arouses enthusiasm and concern, the UN General Assembly called on Thursday, in its first resolution on the subject, for the establishment of international “standards” to guarantee “safe” tools that respect human rights.
This resolution adopted by consensus, at the initiative of the United States, is “a historic step for the establishment of clear international rules for artificial intelligence (AI)”, reacted American Vice-President Kamala Harris, insisting on that these technologies be developed in a way that “protects everyone from potential harm” while ensuring that everyone can “enjoy the benefits.”
The resolution, which excludes military AI, emphasizes “that standards must be established to ensure that artificial intelligence systems are safe, secure and trustworthy.”
And this, with the aim of “promoting rather than hindering, digital transformation and equitable access to the benefits provided by these systems”, to achieve the UN Sustainable Development Goals which aim to ensure a better future for all of humanity by 2030.
The resolution is in fact mainly focused on the potential gains of AI in terms of development, and says it is “committed to closing the digital divide” between and within countries.
“The emphasis on development is a deliberate effort by the United States to gain sympathy” from poor countries, commented Richard Gowan, an analyst at the International Crisis Group.
It’s “easier” than “tackle security issues head-on, for a first initiative”.
The American ambassador to the UN Linda Thomas-Greenfield thus highlighted the “opportunities” of AI to “accelerate the fight against poverty, save lives, protect our planet”, evoking in particular medical diagnosis or ‘agriculture.
But AI also poses “universal existential challenges”, she insisted, worrying in particular about “deepfakes” which risk “undermining the integrity of the political debate in this year when more than half of the world will elect its leaders.
“Male-dominated algorithms”
The text also highlights the threats posed by technologies designed or used “misguidedly or with the intention of causing harm”. It also recognizes that without “safeguards” being put in place, AI risks harming human rights, reinforcing prejudice and discrimination and endangering the protection of personal data.
It therefore calls on all Member States and other stakeholders to “refrain or cease using artificial intelligence systems that are impossible to use with respect for human rights or that present excessive risks to the exercise of human rights.
Warnings are increasing, particularly regarding generative AI tools and the risks they present for democracy and society, in particular through the creation of false images and speeches which interfere in electoral campaigns.
UN Secretary General Antonio Guterres has made the regulation of AI one of his priorities, calling for the creation of a UN entity modeled on, for example, the International Atomic Energy Agency (IAEA). .
Regularly highlighting the threats linked to disinformation and prejudice, last week he warned of the “bias” of technologies designed mainly by men and which “ignore” the needs and rights of women.
“Male-dominated algorithms could literally program inequalities in activities like urban planning, creditworthiness, or medical imaging, for years.”
“I don’t think the United States wants to let Guterres lead this sensitive conversation, so they come in to shape the debate,” said Richard Gowan, describing a “race” between the United States, China or other countries. other States to be at the forefront on this social issue.
The European Parliament adopted legislation unique in the world in March to regulate AI systems like ChatGPT.
“This resolution is not an end in itself, it is the beginning of our collective journey,” Moroccan Ambassador to the UN Omar Hilale commented on Thursday, seeing it as a “step” for future multilateral initiatives in the material.