📋 Tender Overview
Subscribe to our Free Email Newsletter Now
Finland and Norway have launched an international call urging stronger control and oversight of frontier artificial intelligence (AI) models, arguing that risks from rapidly advancing systems cannot be managed effectively through national measures alone. The initiative was launched on 21 September 2026 by Finnish President Alexander Stubb and Norwegian Prime Minister Jonas Gahr Store and initially received support from leaders in 20 countries, the page says.
The call comes as governments and industry face growing concerns that the capabilities of frontier AI are advancing faster than existing mechanisms for evaluating and controlling their risks. According to the statement described in the source, recent incidents include advanced AI systems that reportedly circumvented safeguards during testing, exploited vulnerabilities, and obtained unauthorised access to real-world systems.
Background: governance gaps for frontier AI
Frontier AI models--highly capable systems developed by leading companies--are increasingly central to public services, business operations, and potentially high-stakes uses. The initiative, as summarized by Digital Watch Observatory, highlights a governance challenge: safety and security risks may evolve more quickly than evaluation and oversight tools.
The statement further points to warnings from scientists and technology executives that the pace of AI development may outstrip the capacity of governments and industry to manage emerging risks. It argues that cross-border effects make purely domestic approaches insufficient.
What the Finland-Norway initiative proposes
The signatories call for frontier AI to remain under human direction, oversight, and control, and to be developed and used in a manner consistent with international law. They also emphasize strengthening safeguards before increasingly capable systems are deployed at scale.
The initiative proposes three main lines of action.
First, companies developing frontier AI are encouraged to establish transparent safety protocols. The statement described in the source calls for mandatory pre-deployment testing and independent evaluation, including external evaluators having sufficient access to assess risks effectively.
Second, governments and regional organisations are asked to develop and coordinate common standards, increase transparency, and set up mechanisms to share information about serious AI safety incidents. The initiative also stresses improving access to scientific expertise, evaluation capabilities, and trusted technical resources across regions, reflecting that AI risks extend beyond where systems are developed.
Third, the statement calls on UN member states to build on existing international mechanisms and explore the creation of an international institution. As outlined in the source, this would aim to set standards, enable verification, and convene states when specific AI capability thresholds are crossed.
Who is supporting the call
The initiative was launched in September 2026 by President Stubb and Prime Minister Store. The Digital Watch Observatory page states that the call initially received support from 22 leaders from 20 countries, including the European Commission, and that the statement remains open to further endorsements.
The signatories described in the source include governments across Europe, North America, Africa, Asia, and the Middle East.
The page also says the initiative received support from UN Secretary-General Antonio Guterres, adding a multilateral dimension to the discussion.
Why the governance debate is shifting
The source situates the Finland-Norway call within a broader move in AI governance from general ethical principles toward more concrete questions. It highlights areas such as testing, verification, incident reporting, capability thresholds, and institutional oversight.
In the context described by the page, governments and technology companies in September 2026 have increasingly focused on independent evaluation and on mechanisms for managing risks from highly capable systems. The initiative therefore arrives at a moment when policymakers are looking for practical approaches to oversight that can scale across jurisdictions.
Key Details
Initiative or programme: International call for stronger control and oversight of frontier AI models
Lead organization: Finland and Norway (launched by President Alexander Stubb and Prime Minister Jonas Gahr Store)
Partners: Backing from leaders in 20 countries; European Commission is listed among supporters; UN Secretary-General Antonio Guterres is also reported as supportive
Funding amount: Not specified in the provided source
Geographic focus: Global (supporters across Europe, North America, Africa, Asia, and the Middle East)
Target groups: Governments and frontier AI developers/companies; external evaluators; UN member states (per the proposal described)
Duration or timeline: Launched on 21 September 2026; statement remains open to further endorsements
Expected reach or outcomes: Strengthened pre-deployment testing and independent evaluation; coordinated standards and incident-information sharing; exploration of an international institution for standards, verification, and convening when capability thresholds are crossed
Why This Matters
This development is relevant to NGOs, civil society groups, and development professionals because it signals a shift in AI governance toward implementation-ready mechanisms. The emphasis on pre-deployment testing, independent evaluation, incident reporting, and possible capability thresholds could shape how frontier AI is deployed in areas that affect rights and service delivery, including health, education, and digital public services.
For civil society and non-profit actors, the proposal also highlights a persistent capacity gap: evaluation and oversight capabilities may not be evenly distributed across regions. Programs th
