Leading Tech Companies Weigh in on EU GPAI Code of Practice
The European Union’s General‑Purpose AI (GPAI) Code of Practice (Code) is a voluntary framework designed to help leading AI providers comply with the upcoming obligations under the EU AI Act. Read our announcement of the EU GPAI Code of Practice here.
The Code focuses on:
Transparency: requiring documentation of training data, architecture, and use cases
Copyright: ensuring companies avoid infringing content and honor opt-out
Safety & Security: mandatory for systemic-risk models; includes risk mitigation, incident reporting, and adversarial robustness.
The Code goes into effect on August 2, 2025, ahead of the AI Act’s broader enforcement for GPAI in 2027. Though not legally binding, companies that sign on gain regulatory goodwill and a clearer path toward AI Act compliance. Ahead of tomorrow’s deadline, over 25 companies have signed the Code such as Anthropic, OpenAI, Google, Microsoft, Cohere, and xAI in part. See the full list of signatories of the EU GPAI Code of Practice here.
Leading Tech Companies Committing to the EU Code
Anthropic
Anthropic welcomed the Code, confirming its commitment to fully implement it. In its official statement, the company said:
“We believe the Code advances the principles of transparency, safety and accountability—values that have long been championed by Anthropic for frontier AI development… This collaborative approach—combining regulatory frameworks with flexibility—will be essential for Europe to harness AI's benefits while competing effectively on the global stage.”
Anthropic also reinforced its commitment to working with the EU AI Office and safety organizations to ensure the Code remains effective to emerging technologies and expressed hope that the Code will become a global model for responsible AI development.
OpenAI
OpenAI has signed the Code, viewing this commitment as aligned with its broader safety and transparency practices, such as releasing model system cards, conducting external red-teaming, and following its Preparedness Framework. In its official statement, the company said:
“Signing the Code reflects our commitment to providing capable, accessible, and secure AI models for Europeans to fully participate in the economic and societal benefits of the Intelligence Age…[W]e will continue to iteratively improve our approach to safety to help ensure that our technology is used to benefit everyone responsibly, wherever they are in the world.”
OpenAI emphasized that this moment marks a shift in Europe’s AI trajectory from a focus dominated by regulation to one that empowers innovation. The company framed the Code of Practice as an opportunity to “flip the script” and enable European innovators and builders to shape the future with confidence and clarity.
Google has signed the Code, calling it an important milestone in aligning frontier AI development with European values. In its statement, Kent Walker, President of Global Affairs at Google noted:
“We are committed to working with the AI Office to ensure the Code is proportionate and responsive to the rapid and dynamic evolution of AI. And we will be an active voice in supporting a pro-innovation approach that leads to future investment and innovation in Europe that benefits everyone.”
Google emphasized its support for a proportionate and innovation-friendly approach, while noting concerns over some aspects of the Code, particularly regarding how to implement certain transparency and copyright requirements without the “risk of slowing” Europe’s development and deployment of AI.
Microsoft
Microsoft has signed the Code, emphasizing the company’s commitment to working with the AI office to ensure the code evolves based on real-world applications and international best practices. In her statement, Nanna-Louise Linde, Vice President of European Government Affairs at Microsoft noted:
“We are choosing this path to further build trust in Microsoft AI models, support the European AI ecosystem, and demonstrate our long-standing compliance with EU law, while recognizing that the AI Act is a complex regulation that would benefit from simplification.”
Microsoft urged EU policymakers to use upcoming initiatives, such as the digital simplification package and the Forum for Signatories, to refine the Code and the AI Act into clear, practical frameworks that enable European AI competitiveness on the global stage.
Cohere
Cohere has signed the Code, stating it reflects the company’s commitment to working with governments on their goals for responsible AI development and to our customers in the EU. In its official statement, the company said:
“This is the latest step in our commitments to responsibly developing safe and secure AI technologies and being a key partner to enterprises operating in the EU…Our expertise in data security, on-premises deployments, and customization allows us to develop strong collaborations with government and business partners in Europe and beyond, and uniquely positions us to support their sovereign AI goals.”
Cohere emphasized the need for a regulatory framework that balances clarity with flexibility, acknowledging that the Code must remain adaptable to the rapidly evolving nature of AI technologies, diverse use cases, and a wide range of organizational models.
xAI’s Half-Step: Endorsing Safety, Rejecting Regulation
xAI
xAI, founded by Elon Musk, has pledged that it will only sign the Safety & Security chapter of the Code. In its official statement, the company affirmed its commitment to AI safety but raised sharp concerns about the broader framework:
"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"
xAI’s partial endorsement highlights the company’s support for technical safeguards and risk mitigation, while also signaling strong resistance to provisions they view as stifling innovation; particularly those related to transparency and intellectual property.
In response to xAI’s decision to sign only the Safety & Security chapter of the Code, the European Commission clarified that partial signatories are still subject to the AI Act’s broader requirements. The official EU website, including all signatories, stated:
“xAI signed up to the Safety and Security Chapter; this means that it will have to demonstrate compliance with the AI Act’s obligations concerning transparency and copyright via alternative adequate means.”
This underscores the EU’s position that participation in the Code, while voluntary, is a streamlined path to compliance. Companies opting out of certain sections must still meet the same legal standards through other verifiable frameworks or risk noncompliance.
Breaking from the Pack: Meta’s Lone Rejection of the Code
META
Meta has publicly declined to sign the Code, issuing the lone rejection among all major AI developers. In his statement, Joel Kaplan, Chief Global Affairs Officer at Meta criticized the Code as overreaching and legally ambiguous:
“Europe is heading down the wrong path on AI. We have carefully reviewed the European Commission’s Code of Practice for general-purpose AI (GPAI) models and Meta won’t be signing it. This Code introduces a number of legal uncertainties for model developers, as well as measures which go far beyond the scope of the AI Act.”
Meta also aligned itself with a coalition of over 40 European companies that recently urged the Commission to “Stop the Clock” on implementation, warning that the current framework risks stalling the development of frontier models and stifling innovation across the EU’s tech ecosystem. The company expressed concern that the Code could harm businesses building on AI foundations by introducing compliance burdens that exceed the legislative mandate of the AI Act.
Shaping Global Norms: U.S. Companies, European Rules
The rollout of the EU GPAI Code of Practice marks a pivotal moment in Europe’s AI regulatory landscape. While many leading AI developers including Anthropic, OpenAI, Google, Microsoft, and Cohere have embraced the Code as a meaningful step toward transparency, safety, and regulatory alignment, others like Meta and xAI have pushed back, citing concerns over legal overreach and innovation risk. These divergent responses underscore the broader tension between regulatory certainty and technological agility.
What’s especially notable is the dominant role of U.S. companies in shaping this European initiative. U.S.-based AI developers comprise the majority of early signatories, positioning. At the same time, Meta’s sharp rejection highlights the limits of voluntary frameworks and the growing divergence in regulatory philosophy among U.S. firms.
The recent release of the AI Action Plan signals the U.S. government’s position on a largely deregulatory approach; favoring voluntary commitments, executive orders, and agency guidance over binding federal law. As a result, U.S. companies must navigate a fragmented landscape of international and state-level rules, often aligning with frameworks like the EU Code to help shape emerging global norms. Their engagement with the EU reflects both recognition of Europe’s regulatory influence and a strategic effort to help steer the direction of global AI governance.
As the EU AI Act’s full enforcement looms on the horizon, the Code will serve as both a proving ground and pressure test for how voluntary frameworks can support responsible innovation without stifling it. Whether it becomes a global model or a regulatory cautionary tale will depend on how the EU balances enforcement with flexibility, and how global players, particularly from the U.S., choose to engage.
At LegisLogiq, we help organizations navigate the fast-moving world of A.I. regulation with clarity, creativity, and foresight. Whether you’re exploring policy compliance, advocacy, or looking to redefine your A.I. strategy, our team is here to help. From messaging guidance to partnership opportunities, contact us and someone from our team will connect with you.
