Google says it will sign EU’s AI code of practice

Google Commits to Sign EU’s General Purpose AI Code of Practice

Google has officially announced its decision to sign the European Union’s general purpose AI code of practice, a voluntary framework designed to assist AI developers in implementing processes and systems to adhere to the bloc’s AI Act.

In contrast, Meta recently declared that it would not sign the code, criticizing the EU’s implementation of AI legislation as “overreach” and claiming that Europe was on the wrong track regarding AI development.

Google’s commitment comes just days before the rules for providers of “general-purpose AI models with systemic risk” come into effect on August 2. Major companies affected by these rules include Anthropic, Google, Meta, and OpenAI, among others. They will have a two-year grace period to fully comply with the regulations outlined in the AI Act.

In a blog post penned by Kent Walker, president of global affairs at Google, he acknowledged that the final version of the code of practice was an improvement over the EU’s initial proposal. However, he expressed reservations about the AI Act and the code, citing concerns about potential hindrances to Europe’s AI development and deployment.

By signing the EU’s code of practice, AI companies agree to adhere to a set of guidelines, including providing updated documentation on their AI tools and services, refraining from training AI on pirated content, and honoring requests from content owners to exclude their works from AI datasets.

The EU’s AI Act introduces a risk-based regulatory framework for AI applications, prohibiting certain “unacceptable risk” use cases such as cognitive behavioral manipulation and social scoring. Additionally, the act identifies “high-risk” uses like biometrics and facial recognition, as well as AI applications in education and employment sectors. Developers are required to register AI systems and meet risk- and quality-management obligations under the legislation.