Key points
- The EU will begin enforcing new AI transparency obligations from Aug. 2.
- AI-generated content, including deepfakes, must be clearly labelled.
- Providers of advanced AI models will face stricter compliance and risk management requirements.
Main story
The European Union (EU) will expand enforcement of its Artificial Intelligence (AI) Act from Aug. 2, introducing new transparency requirements for AI systems operating across the bloc.
Under the new phase of implementation, AI systems such as chatbots must clearly inform users when they are interacting with artificial intelligence rather than a human.
The European Commission also requires AI-generated or manipulated content, including deepfakes, images, videos and audio recordings, to be clearly identified through visible labels and machine-readable markings to improve transparency and reduce the risk of deception.
The expanded rules place additional obligations on providers of general-purpose AI models, with the EU AI Office overseeing advanced systems that could pose systemic risks.
Developers will be required to maintain technical documentation, implement copyright compliance policies and provide detailed summaries of the data used to train their AI models.
They must also assess and mitigate risks related to cybersecurity, harmful manipulation, loss of human oversight and potential violations of fundamental rights.
The latest enforcement phase also covers prohibited AI practices, including systems that manipulate users, exploit vulnerable individuals or engage in certain forms of social scoring.
The Commission said enforcement will be coordinated by the EU AI Office, national regulatory authorities and the European Data Protection Supervisor.
It added that more than 180 organisations have already joined a voluntary code of practice to support compliance with the new transparency obligations.
Some requirements for high-risk AI systems have been deferred under the AI Omnibus package, with certain provisions now scheduled to take effect from December 2027 and others from August 2028.
The issues
The EU AI Act is the world’s first comprehensive regulatory framework for artificial intelligence. As AI tools become more widely used across industries, the new rules seek to improve transparency, strengthen accountability and reduce risks associated with misinformation, privacy breaches and harmful AI applications while providing greater certainty for developers and businesses.
What’s being said
“The latest phase of implementation would require AI systems such as chatbots to inform users when they are interacting with artificial intelligence instead of a human.” — European Commission.
“The new rules also require AI-generated or modified content, including deep fakes, images, videos and audio recordings, to be clearly identified through labels and machine-readable markings.” — European Commission.
“The measures were designed to combat deception and manipulation while improving public awareness of AI-generated content.” — European Commission.
What’s next
The new transparency requirements take effect from Aug. 2, while regulators will continue rolling out the AI Act in phases, with additional obligations for high-risk AI systems becoming enforceable between 2027 and 2028.
Bottom line
The EU is tightening oversight of artificial intelligence by requiring greater transparency from AI providers, a move that could influence AI regulation and industry standards well beyond Europe.
 will expand enforcement of its Artificial Intelligence (AI) Act from Aug. 2, introducing new transparency ...){kind=link}