There's a regulatory deadline six days from now that most small business owners haven't planned for. On August 2, 2026, the EU AI Act's Article 50 transparency obligations go live — and despite months of headlines about delayed "high-risk" rules, this piece of the law wasn't pushed back. If your business uses a chatbot, runs an AI-generated content workflow, or sells to customers in Europe, you are in scope.
Here's what happened with the timeline confusion. In June 2026, the European Parliament passed the Digital Omnibus on AI, which pushed the high-risk AI system compliance deadline from August 2, 2026 to December 2027. That's the delay most people heard about. What didn't get the same coverage: August 2 is still a hard date for Article 50 transparency requirements, enforcement powers across all member states, and mandatory compliance for general-purpose AI models. The compliance window didn't disappear. It narrowed to a more specific set of obligations — and those obligations touch a much wider set of businesses than "high-risk" ever did.
What actually goes live on August 2? Three things that affect everyday operators:
First, if you deploy any chatbot, virtual assistant, or customer-service AI that interacts with users, you must disclose that the user is talking to an AI at the first point of contact. This is Article 50(1). It doesn't matter if your tool is built on ChatGPT, Claude, or a custom-trained model. The disclosure has to be built into the system, not buried in a terms-of-service page.
Second, if your business produces AI-generated content — images, text, video, audio — that content must be marked as machine-made. If you're a content studio, a marketing agency, or a solo operator using AI to create deliverables for EU-based clients, that marking requirement now applies to new systems. Systems already on the market before August 2 get a grace period until December 2, 2026.
Third, enforcement powers are now active. Fines for violations can reach up to €15 million or 3% of global annual turnover. For a small business, that math is uncomfortable. The good news: the European Commission published specific guidance acknowledging that SMEs have limited compliance resources, and penalties are expected to be applied proportionally in early enforcement. But "proportional" is not the same as "no consequences."
What does this mean if you're a bootstrapped founder or a lean operator?
Start with a one-hour audit. Open each customer-facing tool you use that touches any EU user. Check whether your chatbot, live chat integration, or AI assistant discloses its nature upfront. If you use AI-generated images or copy on an EU-facing site, check whether those assets carry any disclosure. Most of this is fixable in an afternoon, not a compliance project.
If you're using a major vendor, check their documentation. HubSpot, Intercom, Zendesk, and most enterprise chat platforms are building these disclosures into their systems. But "the vendor handles it" is an assumption, not a confirmed fact, until you've verified it.
The bigger strategic point: this is the first of several enforcement cycles, not the last. The high-risk rules coming in December 2027 will require documentation, conformity assessments, and more involved compliance processes for AI used in hiring, performance management, and access to essential services. Building clean habits now — labeling AI-generated content, disclosing AI interactions, keeping a record of which tools you use for what — costs almost nothing today and avoids the scramble when the next deadline lands.
You don't need a law firm to handle August 2. You need an afternoon, a short checklist, and the discipline to verify rather than assume your vendors have it covered. That's the entire job right now.