The development of artificial intelligence (AI) and the associated risks cannot be abandoned “to the whims” of the market, UN experts warned on Thursday. They are therefore calling for tools for international cooperation, without going as far as a global governance agency.
Nearly a year ago, UN Secretary-General António Guterres set up this committee of around forty experts in the fields of technology, law and personal data protection, from academia, governments and the private sector, such as Microsoft, Google-Alphabet and OpenAI.
Their final report, published a few days before the “Summit for the Future” in New York, unsurprisingly notes “the deficit of global governance in the area of AI”, and the virtual exclusion of developing countries from discussions on an existential subject.
Thus, of the 193 UN member states, only seven are part of seven major initiatives related to AI governance (within the framework of the OECD, the G20 or the Council of Europe), and 118 are completely absent, mainly from countries in the South.
However, the very “cross-border” nature of these technologies “requires a global approach,” the committee insists.
“AI must serve humanity fairly and safely,” Guterres reiterated this week. “Left unchecked, the dangers posed by artificial intelligence could have serious implications for democracy, peace and stability.”
Standing Group of Experts on the Issue
In this context, the committee of experts calls on UN member states to put in place tools for better global cooperation, to encourage the progress of humanity and to avoid slippages.
Because “no one” can predict the evolution of these technologies today and those who make decisions are not accountable when developing and using systems “that they do not understand”.
In these circumstances, “the development, deployment and use of such technology cannot be left only to the whims of the markets,” they insist, stressing the “crucial” role of governments and regional organizations.
The first suggested tool is the creation of an international group of scientific experts on AI, inspired by the model of the UN climate experts (IPCC), whose reports are a reference in this area.
These scientists would inform the international community about emerging risks, sectors where additional research is needed or identify how certain technologies could help meet sustainable development goals (elimination of hunger, poverty, gender equality, climate, etc.).
This idea is taken up in the draft Global Digital Compact, still under discussion, which is to be adopted on Sunday by the 193 member states of the UN at the Summit of the Future.
Experts also suggest establishing a regular intergovernmental political dialogue on the subject and a fund to help lagging countries.
Avoid being “caught by surprise”
To serve as a “glue” between the different tools, they advocate the establishment of a light structure within the UN secretariat.
On the other hand, they do not support the idea of a fully-fledged international governance agency, mentioned in particular by António Guterres, on the model for example of the International Atomic Energy Agency (IAEA). Even if their mandate asked them to explore this possibility, “we do not recommend at this stage to establish such an agency”.
“If AI risks become more serious, and more concentrated, it may become necessary for member states to consider a more robust international institution, with powers of oversight, reporting, verification and enforcement,” they temper.
Even though they note that drawing up an exhaustive list of risks is a “lost cause” in a sector with ultra-rapid development whose evolution no one can predict, they nevertheless cite certain identified dangers: disinformation which threatens democracy, more personal deep fakes (particularly sexual), violations of human rights, autonomous weapons, use by criminal or terrorist groups, etc.
“Given the speed, autonomy and opacity of AI systems, waiting for a threat to emerge could mean that it is already too late to respond,” they admit, counting on constant scientific assessment and political exchanges so that “the world is not taken by surprise.”