EU data watchdog expresses concerns over Council of Europe’s AI treaty

Concerns have been raised about potential compromises in human rights standards due to pressure from foreign business interests. The US and other observer nations like Canada, Britain, and Japan have advocated for an opt-in option allowing states to exempt private businesses from certain rules.


With a treaty on artificial intelligence (AI) being negotiated in Strasbourg this week, the European Data Protection Supervisor (EDPS) voiced concern on Tuesday, 12 March, claiming the outcome has strayed from its intended objective.

The ‘Convention on Artificial Intelligence, Human Rights, Democracy, and the Rule of Law’ is presented as the world’s first of its kind. The Council of Europe (CoE), a 46-member international human rights organisation, aims to create a legally enforceable global agreement to protect the CoE’s human rights norms while promoting AI research and development. The European Union’s data protection watchdog, the European Data Protection Supervisor (EDPS), has expressed concerns about the adequacy of the CoE’s current proposal for an AI Convention.

Why does it matter?

The EDPS believes that the proposal does not go far enough in addressing the risks and challenges posed by AI. The text has been significantly weakened from its original version during negotiations at the CoE’s ad hoc committee in charge of the convention. The EDPS described it as a ‘missed opportunity to lay down a strong and effective legal framework’ for protecting human rights in AI development.
The regulatory body cited the debate around limiting the Convention’s scope to public organisations, adding it defies the treaty’s declared policy purpose of being ‘transversal.’

The US, UK, Canada, Israel, and Japan, who serve as observers on the committee, have lobbied to restrict the scope of the convention to public organisations, carving out exemptions for private firms.
The EU attempted to fight this by presenting a text without a default private sector carve-off, as reported in late January. Germany, France, Spain, Czechia, Estonia, Ireland, and Hungary have expressed support for the convention’s limited scope, favouring a broader acceptance.

The most recent draft also excludes technologies created for national security reasons. Last week, dozens of civil society organisations and academics wrote a letter to the CoE, criticising the Convention for giving a ‘free ride’ to digital and security firms. These developments took place the same week the European Parliament enacted the landmark AI Act, making the bloc the first major political organisation to adopt comprehensive legislation to regulate AI technologies.