AI TRANSPARENCY
AI transparency checks
Recognise the interface, inspect visible disclosure, then establish whether the relevant rules apply.
Chat is not always AI
Start by asking the tool owner which mode is enabled. A live-chat provider may connect to a person, use a decision tree or generate responses with AI. Vendor detection alone cannot establish which is in use. This release reports visible signals and asks the owner to confirm functionality.
Article 50 and territorial scope
EU AI transparency duties have applied since 2 August 2026. Applicability depends on facts including territorial scope and provider or deployer responsibilities. Being a UK organisation does not, by itself, establish an EU obligation.
First-interaction disclosure
The report highlights visible AI wording but does not send chatbot messages. Whether a notice is present at the start of a real interaction and meets the applicable duty requires contextual review.
Other AI uses
Synthetic media and public-interest publishing can raise separate transparency questions. This release does not claim to detect deepfakes or infer an editorial workflow from page content.