What Just Happened in the EU?
On August 2, 2026, key transparency obligations from the EU's comprehensive AI Act became enforceable. This isn't the full rollout of the entire act, which has a staggered timeline stretching to 2027 and beyond, but it's a critical first step for any
business that interacts with the public using AI. The new rules, outlined in Article 50 of the Act, are designed to ensure people know when they are dealing with an artificial intelligence system rather than a person. This applies to companies both inside and outside the EU; if your service or its output reaches people in the European Union, these rules apply to you.
The Core Rule: Disclose the Bot
The most immediate and widespread change involves chatbots and virtual assistants. Providers of these systems must now ensure users are clearly informed that they are interacting with an AI. This disclosure can't be buried in the fine print of a terms and conditions document; it must be presented clearly and distinguishably at the start of the interaction. The goal is to eliminate any ambiguity. If a customer is talking to a bot for support, or using an AI agent to complete a task, they have a right to know it's not a human on the other end. This applies not just to text-based chats but to voice assistants and other interactive AI as well.
Beyond Chatbots: Deepfakes and Synthetic Content
The transparency rules extend far beyond customer service bots. Any organisation that deploys AI to generate or manipulate images, audio, or video—content often referred to as 'deepfakes'—must now clearly disclose that the content is artificial. This also applies to AI-generated text on matters of public interest, unless it has undergone substantial human review and a person takes editorial responsibility. Furthermore, providers of generative AI systems that create synthetic content are required to use machine-readable markings to indicate its origin. There's a short grace period for this specific marking requirement for systems already on the market, giving them until December 2026 to comply.
Why Now? The Push for Trust and Accountability
The EU's motivation is to build sustainable trust in AI by empowering users and combating misinformation. As AI becomes more sophisticated, the potential for deception grows. By making disclosure mandatory, the regulation aims to give people the context they need to critically evaluate the information they receive and the interactions they have. It’s a shift from a world where AI can operate invisibly to one where its presence must be announced. To back this up, the newly established EU AI Office has launched tools for individuals to report suspected infringements, adding a layer of public accountability.
The Global Ripple Effect and Penalties
Like the GDPR before it, the EU AI Act has an extraterritorial reach. Indian companies and others outside the EU are subject to these rules if their AI systems or the content they produce are used by people within the EU. The financial consequences for non-compliance are severe. Violating these transparency obligations can lead to fines of up to €15 million or 3% of a company's total worldwide annual turnover, whichever is higher. This strong penalty structure signals that regulators are serious about enforcement and expect global companies to adapt to this new standard of digital honesty.
