The European Union's landmark Artificial Intelligence (AI) Act entered a critical new phase in June 2025, with several key obligations for companies coming into effect. Following the initial ban on "unacceptable risk" AI systems like social scoring in February, a new set of rules targeting General-Purpose AI (GPAI) models is set to apply from August 2, 2025.
Providers of GPAI models, including large language models, must now maintain detailed technical documentation, implement policies to comply with EU copyright law, and publish comprehensive summaries of the content used for training their models. As of June 2025, there is also a requirement for generative AI systems to clearly label their outputs to inform users that the content is AI-generated. The EU AI Act is the world's first comprehensive law regulating AI and has an extraterritorial reach, applying to any company offering AI products or services within the EU, with non-compliance penalties of up to €35 million or 7% of global annual turnover.
Source: Indeed