{"slug": "chatbots-must-now-confess-they-re-ai-and-deepfakes-need-labels-under-new-eu", "title": "Chatbots Must Now Confess They're AI and Deepfakes Need Labels Under New EU Rules", "summary": "Chatbots operating across the European Union must now tell users they are talking to a machine, and AI-altered images, video, and audio require clear labelling, under new transparency rules that took effect on Sunday. The obligations form part of the bloc's Artificial Intelligence Act, which the Council of the EU has called the world's first law on artificial intelligence. From 2 August, the European Commission's AI Office, working alongside national authorities, also gained the power to enforce the rules, including the ability to compel documentation, run technical evaluations, and issue fines based on a percentage of a company's global turnover.", "body_md": "# Chatbots Must Now Confess They're AI and Deepfakes Need Labels Under New EU Rules\n\n## Understanding the EU's AI Act and its impact on transparency and enforcement\n\nChatbots operating across the European Union must now tell users they are talking to a machine, not a person. AI-altered images, video and audio also require clear labelling, under new transparency rules that took effect on Sunday.\n\nThe obligations form part of the bloc's Artificial Intelligence Act, which the Council of the EU has called the world's first law on artificial intelligence. From 2 August, the [European Commission's AI Office](https://www.ibtimes.co.uk/us-state-department-generative-ai-playbook-1811164), working alongside national authorities, also gained the power to enforce it.\n\n## What Chatbots and Deepfakes Must Now Disclose\n\nUnder the transparency provisions, interactive AI systems such as chatbots must disclose that users are dealing with AI rather than a human. [Deepfakes](https://www.ibtimes.co.uk/minnesota-ai-nudification-law-challenge-1811851), defined as images, video or audio that have been edited or generated using AI, must be labelled as such.\n\nAI-generated or altered content must also carry machine-readable marks, making it easier for platforms and regulators to detect. The Commission says the aim is to reduce deception and manipulation and help people make informed choices online.\n\nMore than 180 organisations have already signed the Commission's Code of Practice on transparency of [AI-generated content](https://www.ibtimes.co.uk/elon-musk-spacexai-challenges-minnesota-ai-nudification-ban-1811380), a voluntary framework designed to help companies show they are meeting the new obligations.\n\n## Fines and Technical Inspections Now Active\n\nSunday's deadline is significant for a second reason. According to the European Commission's AI Act Service Desk, the AI Office and national authorities formally gained the power to enforce the rules on prohibited AI practices, transparency and general-purpose AI models from 2 August, rather than simply monitoring compliance.\n\nThat authority had existed on paper for general-purpose AI providers since August 2025, but the ability to compel documentation, run technical evaluations or issue fines only became active this weekend. Fines under the Act are based on a percentage of a company's global turnover from the previous year, or a set amount, whichever is higher.\n\nProportional penalties apply for smaller firms.\n\nNot every rule arrived on schedule. The transparency and enforcement provisions took effect as originally planned, but not all of the Act's toughest requirements did. According to the Commission's own regulatory framework page, rules for high-risk AI systems embedded in regulated products, such as machinery, have been pushed back to 2 August 2028.\n\nSeparately, high-risk use cases in sensitive areas including recruitment, credit scoring and law enforcement now face a deadline of 2 December 2027 rather than this year. Both extensions followed the EU's formal simplification package, agreed in May 2026 and in force from 27 July, just days before the Act's main application date.\n\nThat timing means the systems most likely to affect people's jobs, finances or interactions with the state have more time to comply, while everyday tools like chatbots and AI image generators face scrutiny immediately.\n\n## What the Changes Mean for You\n\nFor the public, the practical change is visibility. Anyone using an AI chatbot for customer service, or scrolling past an AI-edited photo or video, should now see some form of disclosure identifying it as machine-made or machine-altered.\n\nFor businesses, the shift is accountability. Companies that fail to label content or disclose AI interactions can now face real financial consequences rather than a warning.\n\nThe enforcement mechanism is now live.\n\nThe rules mark the first time Brussels can actually penalise breaches of its flagship AI law, not just require compliance on paper. Whether national regulators use that power consistently across 27 member states, and how it interacts with the delayed high-risk provisions, will shape how effective the Act proves to be in practice over the coming year.\n\n© Copyright IBTimes 2025. All rights reserved.", "url": "https://wpnews.pro/news/chatbots-must-now-confess-they-re-ai-and-deepfakes-need-labels-under-new-eu", "canonical_source": "https://www.ibtimes.co.uk/eu-ai-transparency-rules-enforcement-1811944", "published_at": "2026-08-02 19:27:57+00:00", "updated_at": "2026-08-02 19:43:05.846207+00:00", "lang": "en", "topics": ["ai-policy", "ai-ethics", "generative-ai"], "entities": ["European Union", "Council of the EU", "European Commission's AI Office", "Artificial Intelligence Act"], "alternates": {"html": "https://wpnews.pro/news/chatbots-must-now-confess-they-re-ai-and-deepfakes-need-labels-under-new-eu", "markdown": "https://wpnews.pro/news/chatbots-must-now-confess-they-re-ai-and-deepfakes-need-labels-under-new-eu.md", "text": "https://wpnews.pro/news/chatbots-must-now-confess-they-re-ai-and-deepfakes-need-labels-under-new-eu.txt", "jsonld": "https://wpnews.pro/news/chatbots-must-now-confess-they-re-ai-and-deepfakes-need-labels-under-new-eu.jsonld"}}