On Thursday, the Council of Europe's Framework Convention on AI was adopted in Vilnius, Lithuania. It represents a vital step towards bringing the event of AI into line with human rights, democracy and the rule of law.
The first signatories included the United States, the European Union, the United Kingdom and Israel.
Other signatories include Andorra, Georgia, Iceland, Norway, the Republic of Moldova and San Marino. More countries are more likely to take part the approaching weeks and months.
The project was developed over a period of two years with contributions from greater than 50 countries and goals to create a unified global and legally binding framework for AI governance.
A brand new chapter in AI regulation
This treaty is the most recent in a series of efforts to control AI.
In recent months, quite a few agreements have been reached, including the EU AI lawthe G7 agreement of last October and the Bletchley Declaration signed by 28 countries in November 2023.
In May, 16 international technology firms, including Amazon, Google and Microsoft, signed a recent voluntary safety framework.
AI governance initiatives are proliferating, reflecting growing concerns concerning the potential risks of AI.
However, many still doubt whether voluntary commitments are really effective. Are they greater than just feelings?
The Council of Europe Convention goals to unravel this problem in a legally binding treaty. It obliges signatory states to take protective measures against AI-related threats to human rights, democracy and the rule of law.
The treaty provides for responsibility for harmful consequences and ensures that AI respects the appropriate to privacy.
Věra Jourová, European Commission Vice-President for Values and Transparency, described the convention as a blueprint for global AI regulation: “This framework should create trust that AI innovations respect our values.”
The treaty covers using AI in the private and non-private sectors. Organizations must assess the impact of their AI systems on rights and democracy and publish their results.
In addition, users have to know once they are interacting with AI quite than humans and may challenge decisions made by AI.
Although the framework is described as “legally enforceable”, it appears to lack concrete sanctions reminiscent of fines and relies totally on monitoring compliance.
There have been isolated actions against AI firms. This week, for instance, Clearview AI was fined $33 million by a Dutch agency for collecting facial images for its image database.
Nevertheless, it’s an additional step towards harmonizing global AI rules and complements the AI law by introducing enforceable rules outside the EU.
The global reach of the convention is crucial. The Secretary General of the Council of Europe, Marija Pejčinović Burić, expressed hope for “many signatures and rapid ratification”.
For the treaty to enter into force, it have to be ratified by five signatories, including three members of the Council of Europe.
Once in force, it would be accessible to countries worldwide, potentially creating a very global standard for AI governance.
