# AI Firms Sign Macron’s Paris Call for Safety and Security Global leaders and tech executives commit to new AI safety protocols in Paris. - **Source:** [The Next Web](https://brifing.si/source/afe24b20-b275-4c19-b9a6-8bb3c8babae7) - **Published:** 2026-11-10 - **Topic:** [technology](https://brifing.si/c/technology) - **Original article:** - **On Brifing:** https://brifing.si/s/7c845b1d-fc31-48ed-8858-cef89ace0d5b ## New AI Safety Declaration Ten major AI companies, including OpenAI, Anthropic, and Microsoft, signed the Paris Call for AI Safety and Security on November 9. Hosted by French President Emmanuel Macron, the agreement aims to establish shared protocols for managing risks associated with advanced frontier models, emphasizing collective security and responsible development practices. ## Key Signatories and Participants The declaration includes six nations: France, Canada, Finland, Norway, the Netherlands, and Bahrain. Corporate signatories feature Mistral, Hugging Face, Cohere, and Synthesia. While leaders like Sam Altman and Dario Amodei joined via video, notable absences include Google, Meta, and the United States government, which did not sign the document. ## Core Safety Commitments Signatories pledged to develop fail-safe mechanisms capable of shutting down models during training or internal use. They agreed to provide clearer documentation on system capabilities and allow qualified independent experts to conduct safety tests. Furthermore, companies will share early warnings regarding serious incidents and establish secure channels for whistleblowers. ## Risk Assessment Frameworks The initiative targets malicious uses such as cyberattacks and weapons development, alongside risks of AI acting independently. Participants aim to publish non-sensitive test results in a common format. This effort builds upon previous safety commitments made in Seoul and discussions held during the G7 summit hosted by France in Evian. ## Criticism Regarding Enforcement French safety organization CeSIA criticized the declaration for lacking teeth, noting it sets no common risk thresholds for model release. Charbel-Raphaël Ségerie argued that public authorities should hold the power to suspend high-risk models. The agreement currently lacks specific deadlines and a fully developed framework for independent system access. ## Focus on Child Safety President Macron extended the conversation to AI's impact on children, advocating for mandatory pre-certification of tools targeting young users. Comparing AI safety to toy regulations, he emphasized that chatbots can influence emotions and mental health. A dedicated charter on safer AI for children was planned for the following day. ## Expert Panel Recommendations During the forum, Macron received a government-commissioned report titled Protect without preventing. The document, prepared by a panel of experts, examines the intersection of generative AI and human vulnerability. This report serves as a foundational analysis for future French policy, aiming to balance technological innovation with necessary societal protections. ## Future Policy Implications The Paris Call represents a shift toward greater oversight of frontier AI, following an initiative launched by Norway and Finland in September. While the declaration remains open for new signatories, its long-term effectiveness depends on establishing concrete enforcement mechanisms and achieving broader international consensus among major global AI powers. ## Sources covering this story - https://thenextweb.com/news/paris-call-ai-safety-openai-anthropic-cesia — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._