EU AI Act (chatbot transparency)

The AI Act sorts AI systems by risk. A few practices are prohibited outright, such as manipulative techniques that cause significant harm. High-risk systems — in areas like recruitment, credit scoring or critical infrastructure — face strict requirements. General-purpose AI models have their own obligations. Many other systems, including customer service chatbots, mainly face transparency obligations.

The Act entered into force on 1 August 2024 and applies in stages. Prohibitions have applied since 2 February 2025, rules for general-purpose AI models since 2 August 2025, and most of the rest, including the Article 50 transparency obligations, since 2 August 2026. The Digital Omnibus amendments postponed some high-risk deadlines to December 2027 and August 2028, but left the Article 50 duty for chatbots unchanged.

Article 50(1) requires providers to design AI systems that interact directly with people so that those people are informed they are interacting with an AI system, unless this is obvious to a reasonably well-informed, observant and circumspect person given the context. The European Commission's guidance says this exception should be read restrictively. Under Article 50(5), the information must be given in a clear and distinguishable manner, at the latest at the time of the first interaction, and meet accessibility requirements.

Article 50 also covers AI-generated content: synthetic audio, images, video and text must be marked as artificially generated in certain cases, and deepfakes must be disclosed. Fines for breaching these obligations can reach 15 million euros or 3% of worldwide annual turnover, with proportionality for SMEs.

The provider is whoever develops an AI system, or has it developed, and places it on the market under its name. The deployer is the organisation that uses it in a professional context — for example a store that installs an AI assistant on its website. For chatbots, the Article 50(1) duty sits with the provider, but in practice the disclosure appears in the deployer's chat window, so both have an interest in getting it right.

A simple, effective approach: name the assistant as an AI, state it in the first message or next to the chat window, and avoid presenting it as a human employee. This page is general information, not legal advice.

With AI SmartTalk, the business chooses its assistant's name and writes its instructions, so it can make the assistant introduce itself as an AI and check that this is visible from the first exchange on each channel. The human handoff lets customers reach a person, and EU-hosted models are used by default. Businesses with specific questions about their obligations should seek legal advice.