Despite pressure from some EU countries, the European Commission is still trying to prevent the exclusion of private companies from the first international AI treaty.
The Council of Europe, an international human rights body with 46 member states, is close to finalizing the Convention on Artificial Intelligence, Human Rights, Democracy and the Rule of Law.
From the beginning, the United States, home to the world's leading AI companies, has sought to exclude the private sector from the contract. Once ratified, it will be binding on the state that signed it.
The United States is not a member of the Council of Europe, but participates in the process as an observer. In other words, Washington has no vote, but it can influence the debate by threatening not to sign the agreement.
The pressure on the United States and other countries in favor of a narrower scope of the treaty was so great that the entire drafting process remained secret. NGOs were excluded from this process, although this contradicted the Council of Europe's internal guidelines on stakeholder engagement.
Broad scope or participation
On the other hand, the European Commission, which represents the European Union in negotiations, spoke out against excluding the private sector. Two weeks ago, Euractiv revealed an internal memo saying that “the union should not agree to alternative proposals that limit the scope of the agreement.”
At a subsequent meeting of the Working Group on Communications and the Information Society, the EU Council's technical body responsible for digital policy, several member states called on the Commission to show more flexibility regarding the scope of the agreement.
In particular, countries such as Germany, France, Spain, the Czech Republic, Estonia, Ireland, Hungary and Romania believe that the treaty aims to achieve global agreement. For this reason, attracting more signatories should take priority over a broad agreement with fewer international supporters.
Since the 27 EU member states are among the 46 member states of the Council of Europe, the EU's position can decisively influence the balance of power within the human rights body, where decisions are made by consensus.
The European Union's global ambitions
Restricting the scope of the Convention would constitute a serious blow to the ambitions of the Global Commission. This contract sees a way to establish the European Artificial Intelligence Law, the world's first comprehensive law on artificial intelligence, as a global standard in this field.
In fact, the Commission's negotiating mandate on behalf of the Union is based on the Artificial Intelligence Act. The Commission has shown little willingness to go beyond the AI regulation, even in areas where there is no direct conflict, although the nature of the two initiatives is significantly different.
As part of alignment with the AI Act, the Commission is pushing for far-reaching exceptions for AI applications in the areas of national security, defense, and law enforcement. If the treaty were limited to public institutions only, there would be few such exceptions left.
Such a relaxation of the AI Treaty after several years of commitment by the countries concerned could also discourage future initiatives in this area.
But the final word on this matter has not yet been said. The committee's original memorandum was distributed in preparation for the plenary meeting of the Council of Europe's Artificial Intelligence Committee, which began on Tuesday (January 23) and continues until the end of the week.
Participating countries are expected to reach consensus on the scope during this plenary session with a view to formally adopting the agreement at ministerial level in May. Discussions, including informal discussions last week, have so far been inconclusive. A final decision is not expected until Friday.
Option to unsubscribe
Ahead of the plenary session, the Commission shared an updated note with national EU delegates stating that “the Union must continue to aim to ensure a comprehensive scope of application of the Convention, covering public and private actors.”
Notably, an added paragraph highlights that “in order to preserve the international nature of the Convention, the European Union may nevertheless be willing to consider the possibility of a Party expressing a reservation and relieving itself of the obligation, subject to certain conditions and limitations: Application of the Convention to private parties that It does not act on behalf of or obtain artificial intelligence systems for public authorities.
The Commission's proposal appears aimed at undermining Washington's argument that the United States is unable to commit to anything beyond its domestic legal framework.
In October, US President Joe Biden signed an executive order providing a framework for federal agencies to acquire and use AI tools safely and responsibly. The committee's proposal therefore also applies to companies that do not work with the public sector.
More specifically, the Commission proposes a time-limited “opt-out” option that can be revoked at any time, and provides some guarantees that this option will not be abused. This approach would be the exact opposite of what the US government has proposed. It wants to essentially exclude the private sector from the list and give signatories the option to opt-in.
However, the original “opt-in” option was intended to spare the US government the embarrassment of having to exclude private companies from a human rights treaty. It is also understood that Israel and Japan will not sign the agreement if an opt-out option is included in the final text. On the other hand, the United Kingdom and Canada will follow the American decision.
[Bearbeitet von Nathalie Weatherald]
“Alcohol buff. Troublemaker. Introvert. Student. Social media lover. Web ninja. Bacon fan. Reader.”