Regulation
The EU AI Act's Article 50 transparency rules became enforceable on 2 August, and unlike the high-risk provisions that got pushed to 2027, the disclosure duties covering AI chatbots and generated content did not move
Firms offering EU clients an AI trading assistant, portfolio chatbot, or AI-generated market commentary now have a live legal obligation to disclose that a person is interacting with AI, separate from and unaffected by the broader high-risk AI Act deadlines that the Commission's Omnibus package deferred.
The AI Act has spent most of 2026 in the news for the provisions that got delayed. The Commission's Omnibus package pushed the high-risk system obligations under Annex III, the ones covering things like credit-scoring models and the conformity assessments that go with them, out toward late 2027, and most coverage understandably focused on that reprieve. What got less attention is that Article 50's transparency duties were never part of that deferral. They sat on the original schedule the whole time, and 2 August 2026 is the date they became enforceable.
For trading platforms specifically, Article 50 is a narrower but sharper obligation than the high-risk regime. It does not ask a firm to run a conformity assessment or register a system in an EU database. It asks for something much more basic: if a client is talking to an AI chatbot, an AI-generated research summary, or a synthetic voice on a support line, the firm has to say so, clearly and up front. That sounds trivial until you count how many products now sit on the wrong side of that line without having been built with disclosure in mind. AI portfolio assistants, robo-advisor chat interfaces, and auto-generated market commentary distributed to retail clients are all now squarely inside scope if they touch an EU natural person.
The enforcement mechanism is where this gets real rather than theoretical. Unlike the high-risk regime's phased conformity process, Article 50 obligations attach the moment the system is in use, with no equivalent grace period being floated. A US or UK brokerage serving EU retail clients through an AI-branded chat feature does not get to wait for a future compliance sprint; the obligation is live now, and the honest state of most firms' interfaces is that AI disclosure was treated as a UX nicety rather than a legal requirement until this became unavoidable.
The broader pattern worth watching is that the AI Act's rollout is turning out to be uneven by design rather than by accident, hard requirements landing early where the harm is judged easy to name (a person should know they're talking to a machine) and soft requirements sliding later where the harm is harder to specify (what counts as adequately assessed high-risk credit scoring). For trading and brokerage platforms, that means the disclosure obligations arrived first and the deeper model-governance requirements arrive later, which is the opposite order from how most compliance teams built their 2026 roadmaps.
Read the original: Holland & Knight / Gibson Dunn legal analysis - The EU AI Act's Article 50 transparency rules became enforceable on 2 August, and unlike the high-risk provisions that got pushed to 2027, the disclosure duties covering AI chatbots and generated content did not move. Commentary is the independent editorial view of Share Trading; the original article is credited to its publisher.