Since August 2, mandatory requirements for general-purpose artificial intelligence (GPAI) systems have come into force in the European Union. Companies are now obliged to publish information about training data, comply with copyright, and ensure the safe development of products. This was reported by the European Commission, according to UNN.
Details
European rules for general-purpose artificial intelligence (GPAI) systems have entered into force.
This will increase the transparency, safety, and accountability of AI systems on the market. These rules will ensure clearer information on how AI models are trained, more effective copyright enforcement, and more responsible AI development.
– the statement says.
GPAI includes models capable of generating language and trained on computational volumes. For the most powerful models with potential systemic risk, additional requirements are provided — in particular, the obligation to inform the European Commission and ensure the safety of the models.
This is the first stage of implementing the European AI Act — a law adopted in June 2024. It categorizes AI systems by risk level and sets stricter requirements for high-risk products, especially in healthcare, justice, and critical infrastructure.
The regulation also introduces mandatory registration for models with a high impact on the public — including those that can create disinformation or violate human rights.
The regulation will fully come into force by 2026, but the first transparency obligations have already become mandatory for companies developing or implementing AI in the EU.
Recall
The Italian Antitrust Authority has launched an investigation into Meta over the artificial intelligence feature in WhatsApp. The company is suspected of abusing its dominant position by installing the feature without user consent.
Ukraine to receive tens of thousands of AI drone kits – FT27.07.25, 18:44 • 10938 views