Last updated:
For each website we loaded the home page once, as a first-time visitor sees it, and looked for a chat widget our scanner recognises. Where there was one, we checked whether the page or the widget, as loaded, told visitors they would be talking to an AI. We did not open the chat, type anything or click anything.
The result for each site is one of two things: an AI disclosure was detected at first view, or no AI disclosure was detected by an automated scan. We publish totals only. No website, company or chat provider is named.
| Measure | Result | Notes |
|---|---|---|
| Popular EU websites in the batch | 2,279 | Top of Tranco list GQG5K, 10 EU country domains |
| Not scanned | 862 | 436 could not be scanned; 33 excluded by robots.txt; 393 not visited because robots.txt was unreachable |
| Chat widget detected by our scanner | 59 | The study sample |
| No AI disclosure detected by an automated scan at first view | 58 of 59 (98.3%) | 95% confidence interval 91–99.7% |
| AI disclosure detected at first view | 1 of 59 (1.7%) |
No country is shown separately: we only report groups of 20 or more sites, and no country reached 20 sites with a detected chat widget.
If your website has an AI chat widget, Article 50(1) of the AI Act requires that visitors are told they are interacting with an AI system, unless that is obvious from the context. Article 50(5) sets the timing and the manner: the information must be given "in a clear and distinguishable manner at the latest at the time of the first interaction or exposure".
A line in your privacy policy or terms is unlikely to meet that. Put the disclosure where the conversation starts: in the chat launcher or the first message, in the visitor's language. See AI disclosure statement examples for wording you can paste in.
To see what your own home page shows at first view, run the free scan. It uses the same scanner as this study. This page is general information, not legal advice.
This is a description of what an automated scan saw, not a legal assessment or legal advice.
In our October 2026 scan, 1 of the 59 popular EU websites with a chat widget our scanner could detect (1.7%) showed an AI disclosure at first view. The other 58 (98.3%) showed no AI disclosure detected by an automated scan before the chat was opened.
No. It reports what an automated scan saw on each home page at first view, without opening the chat. A site may show its disclosure once the chat opens, the chat may be answered by people rather than AI, or the AI may be obvious from context. The study names no site and is not a legal assessment or legal advice.
Article 50(5) of the AI Act says the information must be given in a clear and distinguishable manner at the latest at the time of the first interaction or exposure. For an AI chat widget, the safe place is the chat's opening message or the launcher itself, before the visitor types.
Run the free scan on the homepage. It loads your public page in a real browser, the same way as this study, and shows the most severe finding it detects, along with what it could not check. It is automated technical guidance, not legal advice.
Article50.io is an automated scanner for the Article 50 transparency rules of the EU AI Act. It scans a website for potential Article 50 transparency gaps and generates a report of what the scan found and what it could not check, with remediation guidance, implementation instructions and disclosure wording you can paste in.
The free scan shows your single most severe finding in about 30 seconds — no signup, public pages only.
Automated technical guidance, not legal advice. Citations refer to Regulation (EU) 2024/1689.