Commission urges tech firms to join General-Purpose AI Code ahead of enforcement

The European Commission is intensifying efforts to secure industry support for the Code of Practice on General-Purpose AI (GPAI) ahead of its enforcement on 2 August. A workshop hosted by the Commission’s AI Office is scheduled next week, aiming to address the final version of the Code and highlight the advantages of voluntary compliance for AI model providers. This initiative comes as the release of the Code, originally expected in early May, has been postponed.

The drafting process for the Code began in September 2024, with thirteen appointed experts conducting plenary sessions and workshops to gather industry feedback. However, the process has faced criticism from technology companies, publishers, and rights-holders, who argue that the rules may conflict with EU copyright laws. The US government has also expressed concerns, warning that the Code could hinder innovation, while Meta has publicly declined to sign the current version.

Major US tech firms, including Amazon, IBM, Google, Meta, Microsoft, and OpenAI, have collectively urged the Commission to ensure the Code remains straightforward and avoids unnecessary administrative burdens. Despite these reservations, the European Commission remains committed to publishing the Code before the GPAI provisions of the AI Act come into force in August.

Following publication, the Commission will evaluate companies’ willingness to sign and conduct an adequacy assessment with member states. This could lead to formal adoption of the Code through an implementing act. The AI Act, which regulates AI based on risk levels, began phased enforcement last year, with certain provisions delayed until 2027.