TLDR: Microsoft has committed to the European Union’s voluntary Code of Practice for generative AI, aiming to align with the bloc’s AI Act. This move positions Microsoft as a proponent of responsible AI governance, while other tech giants like Meta have expressed reservations and declined to sign, citing legal uncertainties and concerns about hindering innovation.
Microsoft has formally committed to the European Union’s voluntary Code of Practice for generative artificial intelligence, a significant step towards aligning with the bloc’s comprehensive AI regulations. This decision, confirmed by Microsoft President Brad Smith, underscores the company’s proactive stance on responsible AI development and governance. The Code of Practice, drafted by 13 independent experts, is designed to offer legal clarity to companies navigating the EU’s landmark AI Act, which came into force in June 2024.
Under the terms of the code, signatories are required to publish summaries of the content used to train their general-purpose AI models and implement policies that adhere to EU copyright law. Furthermore, the code mandates the provision and regular updating of documentation on AI tools and services, prohibits training on illegal content, and requires compliance with rights holders’ requests regarding the use of their works in datasets. Microsoft’s engagement with the EU’s AI Office has been described as ‘constructive,’ with the company finalizing commitments to uphold transparency, fairness, and risk mitigation in its AI systems. Brad Smith stated, ‘We believe voluntary commitments, backed by clear legislation, are the path forward for building trust in AI.’ This strategic alignment is particularly relevant as Microsoft’s generative AI tools, such as Copilot and Azure OpenAI, continue to expand across various industries.
Also Read:
- Google Aligns with EU on Generative AI Code, Highlighting Big Tech Regulatory Divide
- Microsoft Negotiates “AGI-Proof” Agreement with OpenAI to Secure Future AI Access
In contrast, Meta Platforms has opted not to sign the EU AI Code of Conduct. Joel Kaplan, Meta’s Chief Global Affairs Officer, articulated concerns that the code introduces legal uncertainties for model developers and includes measures that extend beyond the scope of the AI Act. Kaplan expressed worries that such over-reach could ‘throttle the development and deployment of frontier AI models in Europe, and stunt European companies looking to build businesses on top of them.’ Meta shares these concerns with a group of 45 European companies. Despite this divergence, companies like OpenAI and Mistral have already signed the code, indicating a varied industry response to the EU’s regulatory framework. The AI Act itself is a risk-based framework, prohibiting certain ‘unacceptable risk’ uses like cognitive behavioral manipulation and social scoring, while defining ‘high-risk’ applications such as biometrics and facial recognition in sensitive areas like education and employment. Developers of high-risk AI systems must register them and comply with obligations regarding risk and quality management.
