Stubb and Støre launch international call for control of frontier AI models
UN Secretary-General António Guterres has welcomed the call "A Call for Control of Frontier AI Models," announced in connection with the 81st session of the UN General Assembly.
Stubb and Støre launch international call for control of frontier AI models
UN Secretary-General António Guterres has welcomed the call "A Call for Control of Frontier AI Models," announced in connection with the 81st session of the UN General Assembly.
UN Secretary-General António Guterres on Monday, local time, welcomed the launch of "A Call for Control of Frontier AI Models," led by Finland's President Alexander Stubb and Norway's Prime Minister Jonas Gahr Støre, and urged greater international cooperation on the rapidly developing technology. That is according to a report from the news agency ANI, which is the publicly available source for both the call's content, its list of signatories and Guterres' statements (ANI via MSN). The call demands, among other things, mandatory testing before deployment, independent evaluation and common standards – and, according to the source, has the support of leaders and senior officials from 18 countries, in addition to Commission President Ursula von der Leyen. For Norway, the matter is especially pertinent: according to the report, Norway stands together with Finland at the head of the initiative.
What the call says
According to ANI's account, the call states that the rapid development of frontier AI models carries "serious risks to safety and security if not handled properly." The core position reads, in translation: "AI must remain under human direction, oversight and control. It must be developed and used in line with international law" (ANI).
The specific demands
The call is aimed at three levels, according to ANI's account.
To the companies: Developers of the most powerful models are urged to establish "transparent safety protocols, including mandatory pre-deployment testing and independent evaluation." Qualified evaluators are to be given sufficient access to assess the risks. The demand for access for external evaluators is among the more specific elements, and points toward a field of practice in which the laboratories largely decide for themselves who gets to test the models, and on what terms.
To governments and regional organizations: The call urges them to "further develop and coordinate common standards, strengthen transparency – including shared reporting of serious safety incidents – and ensure that countries in all regions have access to scientific capacity, expertise and reliable evaluation."
To UN member states: They are encouraged to build on existing international mechanisms and "explore establishing an international institution, capable of setting standards, enabling verification and convening states when capability thresholds are crossed." The last of these – a focal point for when models reach certain capability levels – points toward an arrangement that does not exist today, and which would require a decision-making process within the UN system.
Who is behind it
According to the ANI report, the call is supported by leaders and senior officials from 18 countries: Australia, Bahrain, Canada, Denmark, Estonia, Germany, Iceland, Ireland, Kazakhstan, Kenya, Latvia, Moldova, the Netherlands, Singapore, Spain, South Africa, Turkey and the United Arab Emirates – in addition to the President of the European Commission, Ursula von der Leyen. The source does not state what title or role each individual signatory holds, beyond the fact that Stubb and Støre lead the call. The list spans a wide geography, from Europe and North America to the Gulf, Central Asia and Africa.
Guterres' support – and the broader framing
According to ANI, Guterres welcomed the call with reference to the need for greater international cooperation on the rapidly developing technology. "The world cannot afford a race to the bottom on AI safety," he said according to the report, pointing to the Independent International Scientific Panel on AI as a source of knowledge about the technology's opportunities and risks.
At the same time, he welcomed the publication of the panel's first thematic brief, and encouraged external experts – "including from frontier AI labs and safety institutes" – to engage with the panel. This places the call within a larger UN framework in which the panel is to serve as the scientific foundation for member states' governance work.
What remains open
It is worth being precise about what the source material actually documents. All the details about the call – its wording, the list of signatories and Guterres' statements – come via a single ANI-syndicated report, not from the call's primary text or from statements by Norwegian and Finnish authorities themselves. The exact calendar date of the launch and any meeting venue are not stated, beyond the fact that Guterres welcomed the initiative "on Monday, local time" in connection with the General Assembly's 81st session. Nor is it clear from the source whether the call entails binding obligations for the participating countries, or whether it is a purely political declaration – and it should therefore not be characterized either way until the primary source is available.
It remains to be seen whether the call is followed up with concrete negotiations on the proposed international institution, or whether it remains a statement of principles. The main impression, however, is concrete: Norway, with Støre at the forefront, has according to the report taken a visible international leadership role in the question of how frontier AI models should be controlled – and the UN Secretary-General has welcomed the initiative.