Supporting safe and trustworthy artificial intelligence
The Artificial Intelligence (AI) Act, which entered into force on August 1, 2024, sets a global standard for AI regulation. At its core is the aspiration for the EU to lead the way in safe AI by defining trustworthiness requirements for high-risk AI systems.
The JRC played a key role in the entire policy cycle on the AI Act, from conceptualisation to implementation. The JRC has also been collaborating on evaluating the risks of various general-purpose AI models with the Commission’s European AI Office, which is responsible for the implementation and enforcement of the AI Act.
Impact on policy
Informed the positions of all three EU institutions, supporting the entire policy cycle on the AI Act.
Shaping the AI Act was a complex but crucial journey—balancing innovation with fundamental rights. The JRC’s expertise provided essential scientific grounding, guiding us through negotiations and ensuring a future-proof framework. This Act sets a global standard for safe, trustworthy AI.
Supported the entire policy cycle
The JRC actively participated in the full policy cycle on the AI Act, providing early research that contributed to the problem definition and agenda setting. Its neutral scientific support informed the positions of all three EU institutions, and was critical for the trilogue phase of the negotiations.
MEP Brando Benifei (Italy, S&D), co-rapporteur for the European Parliament on the AI Act, explains the JRC’s important role in his quote. The JRC also contributed significantly to the standardisation process underpinning the regulation and is still actively engaged in the implementation phase in collaboration with the AI Office.
Pioneered AI research and analysis
The JRC's AI Watch initiative, launched in 2018 in collaboration with DG for Communications Networks, Content and Technology (CNECT), was the first knowledge service to monitor the development, uptake, and impact of AI in Europe. This initiative was instrumental in framing and formulating the Commission’s first legislative proposal on AI. Other important JRC scientific contributions include research on the social impact of AI and state-of-the-art expertise on algorithmic systems, with its own European Centre for Algorithmic Transparency (ECAT).
Led on standardisation and terminology
The AI Act defines high-level requirements to make high-risk AI systems safe. The act is underpinned by standards that specify the technical means needed to meet these requirements. The JRC undertook the technical steering of standardisation groups to ensure their work is aligned with the needs of the regulation. JRC analysis and feedback on the draft standards clarified the key elements of the regulation from a scientific point of view and enabled experts from industry, civil society and academia to shape their technical contributions to the standards.
The JRC continues to support the development of key standards dealing with AI risk management, AI trustworthiness and cybersecurity, to be ready when the obligations start to apply in August 2026.
In addition to supporting the standardisation aspects, the JRC helped define key terms in the AI Act, with an impact in the scope of the regulation. It also played a strong role in guaranteeing international alignment, for example with the Organisation for Economic Cooperation and Development (OECD).
Engaged with general-purpose AI
The JRC also collaborates with the AI Office on General-Purpose AI. It undertakes scientific research to evaluate a variety of AI models and helps on the classification of the AI models with systemic risks. On a more practical level, the JRC experiments with General-Purpose AI by offering its own staff the opportunity to conduct research on and with generative AI via the GPT@JRC platform. The corporate GPT@EC tool, which was launched in 2024 for all European Commission staff, is based on the successful GPT@JRC platform.
Promoted a holistic approach
Building on its strong longstanding collaboration with DG CNECT, the JRC also provided relevant scientific evidence to more than 9 other Commission departments involved on digital policy. In this way, the JRC helps to facilitate a coherent and holistic approach to achieve a fair digital transition in the EU.