After the States, the companies. Develop artificial intelligence (IA) which serves the common good, “it is not only up to governments to act,” declared Audrey Azoulay, Director General of UNESCO, on February 5, from the Kranj Congress Center in Slovenia. There, on the sidelines of two days of the UNESCO world forum on the subject, the Frenchwoman was pleased to have obtained “the same concrete commitment from large companies”.
Eight companies – Lenovo Group, LG AI Research, Mastercard, Microsoft, Salesforce, GSMA, INNIT and Telefonica – signed an agreement to “build more ethical artificial intelligence”. This is evidenced by the large panel on which their representatives symbolically affixed their signature, under the flashes of photographers and the encouragement of Azoulay.
Without engagement
A little over two years earlier, in November 2021, the 193 UN Member States had already committed to respecting the principles promoted by UNESCO regarding the deployment of these technologies. This recommendation – non-binding – asked them to guarantee the “protection of human rights” when deploying AI.
The UNESCO recommendation will no longer be binding on these eight companies. Concretely, they undertake to implement “risk assessment and mitigation procedures”, before and after the marketing of their AI systems.
“The problem with these voluntary initiatives is that they can collapse when they come into conflict with financial interests,” warns Oliver Marsh, from the NGO AlgorithmWatch, present at this “global forum” in Slovenia. The Briton recalls how the social network “But the UNESCO recommendation could prove to be more robust; we don’t know yet,” he adds.
“It’s up to us to hold them to account!”
“It is up to us, civil society, to demand that these signatory companies be held accountable,” said Vilas Dhar of the Patrick J. McGovern Foundation, a global philanthropic organization based in the United States and also present in Slovenia. Stressing that the average citizen is “not as powerless as he thinks”, he highlights the levers of action within his reach: consumer choices, advocacy or lobbying his government.
This UNESCO forum on the ethics of AI was held the day after the agreement, on February 2, of the 27 EU member states on the draft European regulation of AI, a validation came after intense negotiations aimed at finding the right balance between innovation and security.
