EU AI Act: your chatbot has to say it is AI - and 2 December ends the marking grace period

On 2 August, the EU AI Act came into force, mandating that any product engaging with EU users disclose its AI nature - either before or during the first interaction. This includes chatbots, AI-generated text, images, audio, and video. Deepfakes must also carry a visible disclosure label. If a system was already on the market prior to the Act's implementation, the deadline for completing the disclosure requirements is 2 December 2026.

Non-compliance can result in penalties up to 3% of a company's global turnover. The Act applies to software-as-a-service (SaaS) founders creating chat features, AI content tools, onboarding bots, support agents, and any other product an EU user might interact with. Notably, merely having an EU office does not exempt a company, as the Act follows the user, not the headquarters location.

To ensure compliance, companies should map every AI feature interacting with users, add appropriate disclosure copy at the right point - usually at the first interaction for chatbots or visible labels for deepfakes. Generated content must be marked in a machine-readable way, although the exact technical standard is still being finalized and requires documentation of an interim approach.

Companies are advised to keep records of disclosures for potential regulator inquiries. Shuttle Wright has created a $29 template pack, including an obligation map, adaptable disclosure wording, and a marking checklist, to assist businesses in meeting these obligations. A free obligation checker is also available on their website for those wishing to assess their current compliance status.