OpenAI and Europe: Transforming Responsible AI

The Impact of OpenAI in Europe

Every day, millions of Europeans rely on tools developed by OpenAI for various activities ranging from learning to managing professional tasks. These tools are not limited to individuals but also extend to businesses of all sizes and governments in the region. OpenAI believes that responsible artificial intelligence can play a crucial role in enhancing the competitiveness and economic prosperity of Europe. As the EU's AI Act progresses, OpenAI outlines how it has strengthened its practices regarding safety, transparency, and provenance, in line with European guidelines, and how it plans to continue this evolution.

A Long-standing Commitment to Responsible AI

OpenAI's mission is to ensure that general artificial intelligence benefits all of humanity. This goal involves an ongoing responsibility to maximize the benefits of AI, broaden its access, and manage potential risks. For several years, OpenAI has been developing governance approaches and supporting applicable rules that enable the safe use of AI, a central principle of the EU's AI Act. The rules must be pragmatic, proportionate, and risk-based to foster innovation while ensuring effective governance. In this spirit, OpenAI has contributed to the development of two Codes of Practice: the General Purpose AI (GPAI) Code of Practice and the AI-Generated Content Transparency Code of Practice, both developed through multi-stakeholder processes.

Ensuring the Safety and Security of AI

The GPAI Code establishes a common framework for transparency, safety, and security of general-purpose AI models. OpenAI supports this framework through rigorous internal governance and collaboration with external experts, governments, and partner organizations. For years, OpenAI has thoroughly tested its models before market release, publishes system cards with each major version, and involves external experts in the testing process through its Model Testing Network. The public Model Specification provides insight into how OpenAI shapes the behavior of its models.

OpenAI has also strengthened its underlying governance frameworks. The Preparedness Framework, established in 2023 and updated in 2025, defines how OpenAI identifies, assesses, and manages serious risks associated with advanced AI systems. The Frontier Governance Framework builds on this foundation to align safety and security practices with emerging legal requirements, including the GPAI Code of the EU's AI Act. These frameworks translate the principles of responsible AI into practical decisions regarding risk assessment, protective measures, model reporting, security, incident response, and the involvement of external experts.

Responsible AI governance requires collaboration beyond a single company. Through the Frontier Model Forum, collaborations with the U.S. CAISI and the U.K. AISI, and work on third-party assessment practices, OpenAI supports research on shared safety, external testing, and clearer assessment standards across the ecosystem.

Enhancing Transparency and Trust

OpenAI's support for the AI-Generated Content Transparency Code of Practice is based on years of research and development to improve the provenance of AI-generated media. Users need to have better context about the content they encounter online, particularly whether it has been created or modified by AI. OpenAI's approach to provenance relies on two complementary systems: Content Credentials (C2PA), which provide detailed context, and SynthID watermarks that preserve a signal when metadata does not survive. OpenAI is expanding this work to include audio outputs in addition to images.

In line with its commitments under the Code, OpenAI is working to extend provenance measures for its systems across different modalities, including text, as standards and tools continue to mature. The goal is also to support customers and developers using its models by providing signals, tools, and guidance to meet their own transparency obligations. Provenance remains an evolving area, with challenges such as metadata loss and label transmission across platforms. Therefore, OpenAI advocates for a layered approach and ongoing cooperation across the broader ecosystem.

Cybersecurity: An Example of Governance Adaptation

Cybersecurity is an area where dynamic and practical governance is particularly crucial. AI capabilities that help defenders identify and rectify vulnerabilities can also create new risks of misuse. OpenAI aims to reduce abuse while assisting legitimate defenders in using AI through its Trusted Access for Cyber (TAC) program, thereby enhancing collective resilience.

Since the launch of OpenAI's EU Cyber Action Plan in early May 2026, the company has collaborated with EU and national cybersecurity agencies, private sector partners, and critical infrastructure operators to provide them with the most advanced cyber models. This initiative aims to strengthen cybersecurity resilience across the continent, in line with the European Commission's Cybersecurity and Artificial Intelligence Action Plan, which advocates for a coordinated approach to address the risks associated with advanced AI while harnessing its potential.

Continuing to Work with the Ecosystem

As the implementation of the EU's AI Act continues, OpenAI is reinforcing its compliance approach and learning from regulators and the broader ecosystem. As technology progresses, the rules must remain flexible enough to adapt and allow individuals, businesses, and organizations to fully benefit from the Age of Intelligence.

To support customers and developers preparing for the implementation of the EU's AI Act, OpenAI provides practical resources, including model documentation, system cards, security information, usage policies, and guidance on provenance and verification tools. OpenAI will continue to update these resources as the implementation evolves.