EU AI Act Transparency Rules Now in Force: What They Mean for Gemini, Galaxy AI, and Every Phone With AI Features
The European Union began enforcing its AI Act transparency rules on August 2, requiring every AI system operating in the EU to disclose when users are talking to a machine and to label AI-generated images, audio, and video. The rules carry fines of up to 15 million euros or 3 percent of a company’s global annual turnover — whichever is higher — and they apply to Google, Samsung, Apple, and every other company shipping AI features into European phones.
The practical consequences for Android users are immediate. Gemini, Galaxy AI, and any on-device tool that generates or edits images, video, or audio must now carry visible labels and embed machine-readable watermarks so the content can be detected as AI-made downstream. Chatbots and AI agents — including Gemini’s conversational mode — must tell users upfront that they are interacting with an AI system, not a person.
What the Rules Require
Article 50 of the AI Act sets out four transparency obligations that took effect on August 2:
- Chatbot disclosure: Users must be clearly informed when they are interacting with an AI system rather than a real person. This covers chatbots, AI agents, and avatars.
- Deepfake labeling: AI-generated or manipulated images, audio, and video that resemble existing people, objects, places, or events must be clearly and visibly labeled.
- Machine-readable marks: All such content must also include machine-readable watermarks so it can be detected automatically, not just by a human reading a label.
- AI-generated text: Text published on matters of public interest that was generated by AI without human editorial review must be marked as such.
The European Commission has published a set of standardized icons — “AI,” “AI Generated,” and “AI Modified” — that companies can use for visible labeling. Each comes in black, white, and semi-transparent variants, and the Commission has made them freely available with no attribution required.
Grace Period and Enforcement
Generative AI systems already on the market before August 2 get a grace period until December 2026 to integrate machine-readable watermarks. But the basic requirements — chatbot disclosure and visible content labels — apply immediately. Over 180 organizations have signed the Commission’s voluntary Code of Practice on transparency of AI-generated content, which operationalizes the rules in practice.
Enforcement falls to national market surveillance authorities in each EU member state, the EU’s AI Office, and the European Data Protection Supervisor. For EU institutions themselves, the maximum fine is capped at 750,000 euros.
What This Means for Android Users
Google already embeds SynthID watermarks in Gemini-generated images and has signed the EU’s Code of Practice. Samsung’s Galaxy AI editing tools — photo eraser, generative fill, sketch-to-image — will need to meet the same bar for any content that could be mistaken for a real photograph of a real person or place. The obligation is not optional or best-practice: it is law, with real financial teeth.
Whether these labels actually help ordinary users spot AI-generated content is a separate question. A small icon on a forwarded WhatsApp image is easy to miss, and machine-readable watermarks can be stripped. The rules set a floor, not a guarantee — but for Android users in the EU, this is the first time there is a legal requirement that the AI features built into their phones must announce themselves.
Sources: European Commission, EU AI Office Quick Facts
