AI-driven advertising can use what people say to chatbots to infer their interests and circumstances, the ICO warns in a report on UK data-protection rules. The regulator says existing privacy law applies, but conversational AI could make the personal information available to advertisers more revealing.
Watch Desk analysis
What happened
The ICO’s 51-page report, published on 13 November, examines how large language models and agentic AI are changing online advertising. It covers ads alongside chatbot answers, systems that let AI agents interact with advertising platforms, retail product discovery, connected TVs and walled gardens, among other settings.
The report describes a chatbot answering a holiday question while a separate advertising system uses the travel-related signal to select an advert. The ICO says it examined cases where models passed signals such as the apparent intent of a request to advertising systems. A conversation might also disclose sensitive information directly, or allow a system to infer it.
The regulator says the UK GDPR, PECR and the Data Protection Act 2018 already apply. Its report introduces no new legal requirements, but sets out expectations around accountability, meaningful control, transparency and fairness, careful handling of sensitive inferences, supplier checks, data minimisation and risk assessment. Resultsense reports that the findings will inform planned ICO guidance on agentic AI and automated decision-making.
Why it matters
A chatbot conversation can reveal more than a search or a click: it may contain immediate plans, personal circumstances or sensitive details. If those signals travel from a conversational AI product into ad targeting, the privacy question is not just what someone deliberately disclosed, but what an advertising system can infer and how that information is used.
The ICO’s point is that a new interface does not replace the rules already in force. For companies building conversational adverts or agentic advertising tools, the report puts user control, transparency and data handling squarely in the compliance picture.
Our read
This is a useful reminder that “the chatbot answered me” and “the ad system learnt something about me” may be two parts of the same transaction. The ICO is describing how existing rules apply, not announcing a new ban or a new legal test. Its planned guidance should make the practical expectations clearer. Until then, businesses should not treat a conversational prompt as a harmless substitute for a consent form.
What to watch
- The ICO’s planned guidance on agentic AI and automated decision-making.
- Whether firms explain when chatbot conversations or inferred signals inform advertising.
- How the regulator’s Regulatory Sandbox and future monitoring address agentic advertising.
Discussion spark: Should conversational AI products be allowed to use inferred details from a chat for advertising if users are told clearly, or should that use require separate consent?
Sources and evidence
- ICO AdTech 3.0 report: agentic AI and online advertising – Resultsense (16 November 2026, 12:45 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.