Official Lines same event, every official angle

2026-11-09 – 2026-11-10 · 2 statements · 2 actors

Paris Call for AI safety and security (9 November 2026)

Background

A joint declaration titled the Paris Call for AI safety and security was published by France on 9 November 2026 and by Canada on 10 November 2026. According to the text, policy makers, researchers and developers commit to a framework for cooperation on the evaluation and risk management of frontier AI models. Only this joint text is provided, so no other actor's account is available.

Compiled by AI only from the official statements below, not from other sources. Facts no source disputes are stated plainly; everything else is attributed. How this works.

How the official lines relate

Only one voice is present: the joint Canada-France text. It states that no single actor can manage AI risks alone and sets out commitments on evaluation frameworks, publication of findings, fail-safe mechanisms and whistleblower channels. No other actor's statement is given for comparison.

Spoke

Active on this topic recently, no statement on this event

Where the official lines meet and part

● asserts   ◐ qualifies   ○ denies   blank: no stance in the statements we have. Hover or tap a symbol for the quote. Shared floor: every actor in the table asserts it. Uncontested majority: more than half assert it, nobody denies it, the rest are silent or qualify. Contested: one asserts, another denies. Definitions.

PropositionCanada and France (joint text)
Policy makers, researchers and developers should establish a common framework for the evaluation and risk management of frontier AI models. Governance of frontier AI risks uncontested majority ●
Evaluation findings that are not sensitive should be published in a standardized format. Governance of frontier AI risks uncontested majority ●
Developers of frontier models should implement fail-safe mechanisms to shut down model operations within their control. Governance of frontier AI risks uncontested majority ●
Policy makers should offer secure and confidential channels for whistleblowers to raise concerns about systemic AI risks. Governance of frontier AI risks uncontested majority ●
No single actor (State, company or international organisation) can manage the risks of AI alone. Governance of frontier AI risks uncontested majority ●

Canada and France published the same text. It is one column: one text is one voice, however many publish it. Agreement between its publishers is not measured; it is listed under shared wording.

Notable patterns

  • Shared wording: Canada and France publish 4 identical passages, the longest 424 words: “and the absolute necessity of implementing rigorous evaluation standards. This declaration reflects our commitment to making a substantial contribution and accelerating the implementation of these priorities. To operationalize this shared …”

Identical passages can be a joint statement or common talking points. All shared wording across events.

Emphasis

  • Canada and France (joint text): Presents AI governance as a shared global responsibility, emphasising cooperation, common evaluation standards and safety commitments alongside innovation and wide adoption.

The statements

France

  • Paris Call for AI safety and security. 2026-11-09 · head of government

    The signatories state a shared interest in safe, secure and widely available advanced AI and set out three strategic commitments: common evaluation and risk-management frameworks, accountability for all actors, and information sharing and early-warning systems.

Canada

  • Joint declaration on Paris call for AI safety and security 2026-11-10 · head of government

    The declaration, as published by the Prime Minister of Canada, lists signatories including Heads of State and Government (Bahrain, Canada, Finland, France, Norway, the Netherlands), AI companies and civil society organisations. It states it remains open to new endorsements.

Compared 2026-11-13 with claude-sonnet-5-5. How this works.