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

Google has officially announced its intention to adopt the European Union’s AI code of practice, a voluntary framework designed to guide AI developers in aligning with the forthcoming AI Act. This decision comes in stark contrast to Meta, which recently rejected the code, labeling the EU's regulatory efforts as excessive and asserting that Europe is misguided in its AI approach. The timing of Google’s commitment is significant, as the new regulations regarding 'general-purpose AI models with systemic risk' will take effect shortly on August 2. Key players in the AI landscape, including Anthropic, Google, Meta, and OpenAI, are among those who will be impacted by these regulations, which mandate a two-year compliance period with the AI Act. In a blog post published Wednesday, Kent Walker, Google’s president of global affairs, acknowledged improvements in the final draft of the code compared to earlier proposals. However, he expressed ongoing concerns regarding the AI Act and the associated code, cautioning that they could hinder AI innovation in Europe. Walker pointed out potential issues such as departures from existing EU copyright laws, delays in approval processes, and mandates that could compromise trade secrets, all of which might stifle the development and deployment of AI technologies in the region. By signing the EU’s code of practice, AI companies agree to adhere to various guidelines, including the obligation to provide updated documentation about their AI tools, refraining from training AI on pirated content, and respecting content owners’ wishes regarding the use of their materials in datasets. The EU’s AI Act is characterized by a risk-based regulatory approach, banning certain high-risk use cases, such as cognitive behavioral manipulation and social scoring, while also establishing definitions for high-risk applications involving biometrics, facial recognition, and sectors like education and employment. Furthermore, the act mandates that developers register their AI systems and fulfill comprehensive risk and quality management requirements.