The European Union has introduced new transparency requirements under the EU AI Act aimed at helping people identify AI-generated and manipulated content and recognise when they are interacting with artificial intelligence rather than a human.
The measures address growing concerns around misinformation, impersonation, fraud and consumer deception as AI-generated content becomes increasingly difficult to distinguish from authentic, human-created material.
Under the new requirements, providers and deployers of certain AI systems will be required to clearly disclose when content has been generated or manipulated using AI. The rules cover images, audio and video that resemble existing people, objects, places, entities or events, including deepfakes.
AI-generated or manipulated content will need to be clearly and visibly labelled, while also incorporating machine-readable markings. The European Commission has also developed a set of icons that can be used to indicate AI-generated content.
The transparency requirements extend beyond synthetic media. They also apply to emotion recognition and biometric categorisation systems, as well as AI-generated text published to inform the public about matters of public interest where the content has not undergone human review or editorial control.
AI systems must disclose when users are interacting with them
The rules also introduce transparency requirements for direct interactions with AI systems.
Users must be clearly informed when they are interacting with an AI system rather than a human. This can apply to technologies including chatbots, AI agents and AI-powered avatars.
The European Commission has published guidelines to help providers and deployers understand and demonstrate compliance with the transparency obligations. Compliance can include adherence to a relevant code of practice.
Fines of up to €15 million
Enforcement of the transparency requirements will involve national market surveillance authorities, the European AI Office for systems under its supervision, and the European Data Protection Supervisor when EU institutions act as providers or deployers.
Companies that fail to comply could face fines of up to €15 million or 3 per cent of their global annual turnover, with proportionality taken into account for small and medium-sized enterprises (SMEs) and small mid-cap companies (SMCs).
EU institutions, bodies and agencies could face fines of up to €750,000.
The EU AI Act entered into force on 1 August 2024, with its provisions being implemented progressively and different obligations taking effect at different stages.
The legislation establishes harmonised rules for AI across the EU, with the broader objective of supporting trustworthy AI development and adoption while addressing risks to health, safety and fundamental rights.
With generative AI increasingly being used to produce realistic images, audio, video and text, the new transparency requirements are intended to give users greater visibility into the origins of digital content and their interactions with AI systems.
The European Commission has identified effective implementation of the AI Act as a key priority as the EU seeks to balance AI innovation with safeguards for democracy, fundamental rights and the rule of law.
