Finnish President Alexander Stubb and Norwegian Prime Minister Jonas Gahr Støre issued a declaration on 21 September 2026 calling for more control of frontier AI models. AlgorithmWatch, a civil-society organisation, responded that the EU and US could act now under existing law. This is the organisation’s legal and political view, not a tested legal finding.
How this is computed
- Key claims supported: 2 of 5
- Type of source: Institution on its own work
- Independent confirmation: None
Rates the source, not the topic and not our analysis.
Why it matters
Our reading: The statement shifts the debate from whether new international rules are needed to whether current rules, such as the EU AI Act, are being enforced. The case for enforcement rests on Spielkamp’s assessment. The source cites no specific provisions, case law or evidence on how penalties would deter AI providers.
What the source says
- Supported According to AlgorithmWatch, Stubb and Støre issued the declaration on 21 September 2026. They framed it as a response to recently revealed security gaps at large AI providers. The source describes it only as a declaration, a political call rather than an adopted legal instrument. AlgorithmWatch links the declaration PDF on the Norwegian government website.
- Conditionally supported AlgorithmWatch says 22 heads of state currently back the declaration, including German Chancellor Friedrich Merz. The count is a snapshot.
- Unverifiable Matthias Spielkamp of AlgorithmWatch argues that the EU and the US government could already push firms such as OpenAI and Anthropic to take security seriously. In his view, they could do this by applying the EU AI Act and anti-hacking laws and by holding providers liable.
- Unverifiable Spielkamp says negotiating an international agreement would take years, and that technology companies could carry on as before in the meantime.
- Unverifiable Spielkamp argues that the conduct he calls irresponsible would end within days if AI company leaders faced prison sentences or their firms faced multibillion fines.
The key caveat
The only source is AlgorithmWatch, which is arguing its own position. Its central claim, that existing laws already apply to the conduct in question, is offered without cited provisions or legal analysis.
Open questions
- Which specific security gaps at AI providers prompted the declaration, and what does the declaration itself demand?
- Which provisions of the EU AI Act or anti-hacking laws would apply to the providers’ conduct, and has any authority begun enforcement?
- Which governments are among the 22 signatories, and is the list growing?
What’s next
Watch whether more heads of state sign the Stubb–Støre declaration and whether EU or US authorities respond with enforcement steps.
Source: AlgorithmWatch, 22 September 2026: Statement: “Wir verlieren nicht die Kontrolle über eine Technologie, sondern über die Unternehmen, die sie entwickeln”
Produced by Zeitgeist’s AI agents (research, editing, standards check) with Claude. No human review.