AI Labels Compulsory on Authentic-Looking Content Under EU Rules
From 2 August, the AI Act’s transparency rules require deepfakes, synthetic media, and machine-written text to be marked and disclosed, with fines reaching 3% of global turnover.
Background
July 31, 2026 – 11:22 am
Image by: Canva
Companies that make or use AI will soon have to label content that looks authentic but is not, as the European Union’s landmark AI Act moves from principle into enforcement.
Key Points
- From 2 August: Providers and deployers across the bloc face binding transparency rules covering deepfakes, synthetic media, and machine-generated text.
- Labeling Requirements:
- Providers: Must embed machine-readable markings for AI-generated or manipulated images, audio, video, and text.
- Deployers: Must disclose the origin of such content to the audience in a clear and timely manner.
- No Loophole for Intent: The duty to label realistic deepfakes applies whether or not the creator intended to deceive.
- Chatbots Included: Interactions with AI systems must inform users at the moment of contact, not just in terms of service.
- Carve-outs and Light Touch:
- Clearly unrealistic or fantastical content is exempt.
- AI-generated text escapes labeling if a human with genuine editorial responsibility has reviewed it.
- Creative work, including film and satire, gets a lighter disclosure requirement.
- Penalties: Breaches can incur fines of up to €15 million or 3% of a company’s worldwide annual turnover.
- Timing Adjustments: The obligations formally apply from 2 August, but industry pressure has led to potential delays for some provisions.
- Industry Lobbying: Advertisers and others have lobbied against AI-generated ads being subject to the transparency rules.
- Guidance: The Commission has published draft guidelines and is finalising a code of practice on marking and labelling AI content.
- Global Impact: Because global platforms rarely build separate systems for single markets, the EU’s labelling regime may shape global practices.