xAI to sign ‘Safety Chapter’ of European Union AI code, unclear on rest

Elon Musk's AI venture, xAI, made headlines on Thursday by announcing its commitment to sign the Safety and Security chapter of the European Union's Code of Practice. This framework, which remains non-binding, is intended to guide AI companies in aligning with the EU's forthcoming regulations on artificial intelligence. The Code of Practice was crafted by a panel of 13 independent experts and is divided into three primary sections: transparency, copyright, and safety and security. While general-purpose AI developers are expected to adhere to guidelines concerning transparency and copyright, the safety chapter particularly addresses creators of more complex AI systems. By endorsing this safety chapter, xAI aims to navigate the evolving regulatory landscape within the EU with greater legal assurance. However, the company has voiced significant concerns regarding other segments of the Code. In a recent statement on the social media platform X, xAI asserted, "xAI supports AI safety and will be signing the EU AI Act’s Code of Practice Chapter on Safety and Security. While the AI Act and the Code have a portion that promotes AI safety, its other parts contain requirements that are profoundly detrimental to innovation, and its copyright provisions are clearly an over-reach." At this point, xAI has not disclosed whether it plans to sign additional chapters related to transparency and copyright, and the company has not provided further comments outside of routine business hours. Responses from major tech firms regarding the Code have varied widely. Google has expressed its intention to sign the entire framework, while Microsoft is also likely to follow suit. In contrast, Meta has chosen to opt out, citing concerns over legal ambiguities and suggesting that the Code exceeds the intended scope of the AI Act. Although the EU's AI Code of Practice is voluntary, companies that sign it may benefit from enhanced legal clarity as they prepare for the bloc's forthcoming legally binding AI Act, expected to take effect in the coming years.