EU Commission Receives Final Version of AI Code Amid Industry Protests
Brussels: The European Commission has officially received the final version of the Code of Practice on General Purpose Artificial Intelligence (GPAI) models. This Code, pending approval from the Commission and the EU27 member states, aims to clarify rules within the AI Act that will take effect starting August 2. These rules specifically target GPAI model providers with systemic risks, including OpenAI’s GPT-4, Google’s Gemini, and xAI’s Grok.
According to Ansa News Agency, the Code comprises three main chapters focusing on transparency, copyright, and risk assessment and mitigation. While the transparency and copyright chapters apply to all GPAI models, the risk assessment chapter affects only the most powerful and advanced models. Developed by 13 independent experts with contributions from over a thousand stakeholders, the Code serves as a voluntary tool for model providers, SMEs, academics, AI safety experts, rights holders, and civil society organizations.
Signatories to the Code are expected to experience reduced administrative burdens and increased legal certainty compared to other compliance methods. The AI Office, established within the Commission, has extended implementation deadlines, allowing providers to implement the Code in good faith without being considered in default if they do not immediately fulfill their commitments.
The forthcoming Commission guidelines will further clarify the entities subject to the AI Act’s general-purpose AI regulations. Previous drafts of the Code sparked industry protests due to perceived pressure from Big Tech to dilute the text. Additionally, ongoing negotiations on tariffs have added complexity, with Washington accusing Brussels of unfairly targeting the American tech sector with digital rules.
Despite acknowledging the influence of large developers, Democratic Party MEP Brando Benifei, the rapporteur for the AI Act, emphasized the need for rigorous application, controls, and sanctions for non-compliance. Henna Virkkunen, Commission Vice-President for Technological Sovereignty, Security and Democracy, highlighted the Code’s role in ensuring AI models are both innovative and safe. She encouraged all GPAI model providers to join the Code, offering a collaborative path to compliance with European AI law.