New transparency requirements for artificial intelligence content began to apply across the European Union on 2 August 2026. Companies must inform people when they are interacting with an AI system and make synthetic or manipulated materials identifiable. Businesses that fail to comply may face fines of up to €15 million or 3% of their total worldwide annual turnover.

The obligations are established under Article 50 of the EU Artificial Intelligence Act, commonly known as the AI Act. They apply to providers and deployers of certain generative and interactive systems operating within the European market. Users should therefore be made aware when they are communicating with a chatbot or another automated AI service rather than a human representative.

Providers of generative systems must introduce machine-readable markings that allow synthetic content to be detected. These technical indicators are intended to help platforms, specialised tools and regulators recognise images, audio recordings, videos and text that have been generated or altered using artificial intelligence.

The rules place particular emphasis on deepfakes. Realistic visual or audio materials that imitate existing people, objects, locations or events must carry a clear disclosure explaining that they were artificially generated or manipulated. This is intended to reduce the risk of fabricated content being mistaken for genuine evidence or authentic reporting.

Labelling requirements also cover AI-generated text published to inform the public about matters of public interest when it has not undergone human review or editorial control. This provision may affect automated news stories, public information materials and other publications capable of influencing political or social discussions.

Compliance will be supervised by national market surveillance authorities, the European AI Office and other competent institutions. Penalties will depend on the circumstances of each case, while proportionality may be taken into account when authorities deal with small and medium-sized companies.

The new framework is designed to give European users clearer information about the origin of the digital materials they encounter. Visible disclosures and technical markings should make it easier to distinguish authentic content from synthetic media while encouraging companies to consider manipulation risks when developing AI products. The EU expects the measures to strengthen public protection against deceptive deepfakes and other misleading uses of artificial intelligence.