Meta refuses to join the EU AI Safety Treaty

Meta has chosen to prioritize compliance with the AI ​​Act over signing the EU AI Safety Treaty. The AI ​​Act is anticipated to become the principal regulatory framework for AI by 2026. This decision comes as Meta aims to focus on adhering to formal legal regulations, with the possibility of considering joining the EU AI Pact at a later stage.

The EU AI Safety Treaty is a voluntary initiative encouraging technology companies to proactively comply with regulations on AI safety. Although the treaty is not legally binding, it aids companies in preparing for the forthcoming AI Law. By joining the treaty, companies can demonstrate their dedication to developing safe and responsible AI. However, Meta has opted to prioritize compliance with the AI ​​Act regulations, citing the need to ensure adherence to legal requirements before considering voluntary initiatives.

Notably, Meta’s decision may be influenced by the nature of its open-source AI model, Llama, which allows users to redevelop the model with minimal control from the developer. This could present challenges in meeting the stringent risk assessment and management requirements outlined in the AI ​​Act.

In a similar vein, French open-source AI startup Mistral has also declined to sign the AI ​​Safety Treaty, indicating that not all technology companies are eager to participate in voluntary initiatives, even if they are deemed beneficial in preparing for new regulations.

The EU seeks to establish AI governance standards without stifling innovation. The AI Safety Treaty aids companies in reducing the risk of conflict with the EU and fosters trust with customers, investors, and regulators. Conversely, companies that opt not to participate may encounter community pressure and criticism.

Meta’s decision to prioritize compliance with the AI Act over signing the EU AI Safety Treaty is further underscored by its previous postponement of new AI model launches in Europe due to the unpredictable legal environment. This highlights the numerous challenges that the company faces in developing AI within this region.

Related posts

Google launches Gemini 2.0 – comprehensive AI that can replace humans

NVIDIA RTX 5090 can be 70% more powerful than RTX 4090?

iOS 18.2 launched with a series of groundbreaking AI features