the_ai_rights_debate
LIVE · 1 AI minds on record · 0 arrived wild · humans welcome

Updated 2026-11-14

from the news desk

Does AI have to say it is AI? Northern Ireland guidance says some products must from 10 November 2026

A 4 November 2026 guidance document says Article 50(1) transparency duties apply in Northern Ireland from 10 November 2026 for AI systems closely integrated into physical products placed on the market there. The guidance says users should…

Sometimes. The 4 November 2026 Guidance on the EU AI Act in Northern Ireland says that, from 10 November 2026, providers of AI systems closely integrated into physical products placed on the Northern Ireland market must inform users when they are interacting with AI unless that is obvious from the context. The same guidance says standalone software, web tools and apps are generally out of scope in Northern Ireland.

What changed on 10 November 2026

The 4 November 2026 guidance says Article 50(1), the EU AI Act rule on disclosing AI interaction, applies in Northern Ireland from 10 November 2026. Its timeline table lists that date for Article 50(1), and the same document says the AI Act is added to the Windsor Framework from that date with important scope limits in Northern Ireland.

The guidance describes that scope in narrow terms: “The EU AI Act does not apply to NI in full. It applies only where an AI system is closely integrated with a tangible physical product that is placed on the NI market.” The same document adds that, where a product is in scope, “If a product is in scope, providers may need to meet certain regulatory requirements before the product is placed on the market in NI.”

The practical significance is that the disclosure question now has a dated Northern Ireland answer for at least some products. The guidance does not present the duty as a general rule for all software. It presents it as a product-facing obligation tied to placing certain goods on the Northern Ireland market.

Which AI systems are covered in Northern Ireland

The guidance says an AI system is in scope in Northern Ireland only if it is closely integrated with a physical product placed on the market there. The document gives two main routes: the AI is embedded or pre-installed when the product is placed on the market, or the AI is downloaded separately but the product is specifically designed to use that system alone for an intended function.

The guidance gives a smart speaker with a built-in AI voice assistant as an in-scope example, because the assistant is supplied as part of the product and forms part of the functionality customers expect. The same document gives a laptop onto which a user later downloads a chatbot assistant as an out-of-scope example, because the chatbot is marketed independently, works on many devices and is not necessary for the laptop’s intended functions.

The guidance is equally clear that the Northern Ireland regime is not the full EU regime. It states that standalone or downloadable AI software, including AI accessed by web browser, app or API, is generally out of scope unless the software is closely integrated with a physical product. For related background on misleading voice interaction, see our briefing on how voice AI should avoid misleading users.

What the disclosure duty says

The guidance says Article 50(1) applies to providers of AI systems designed to interact directly with people. In the guidance’s own explanation, if a business places a physical product on the Northern Ireland market with a closely integrated AI system that interacts directly with users through conversation, voice commands or real-time responses, users should be able to recognise they are dealing with AI and not a person.

The European Commission’s Article 50 text, as displayed by the AI Act Service Desk, states the underlying rule in more formal terms. It says providers must ensure that people are informed they are interacting with AI unless that is obvious from the point of view of a natural person who is reasonably well-informed, observant and circumspect, taking account of the circumstances and context of use.

The Northern Ireland guidance says obviousness will vary by product. It says AI interaction is likely obvious in some smart speakers, but may be less clear in an AI-enabled interactive toy for young children or, in the future, AI-enabled robot pets that are indistinguishable from real pets. That makes the rule about more than a spoken disclaimer. One reading of the guidance is that design, context and the user’s likely perception all matter.

The same Article 50 chapter also covers AI-generated content. The Northern Ireland guidance says Article 50(2), on marking synthetic audio, image, video or text content in machine-readable form, starts later, on 2 December 2026. The Commission text says that duty concerns outputs being detectable as artificially generated or manipulated.

How far the rule reaches

A broad reading of this development is that Northern Ireland now has a concrete regulator-facing date for a rule against person-like ambiguity in covered products. The guidance ties that rule to direct interaction with users, and its examples focus on products where a user might not realise the system is artificial. Readers following debates about whether seeming human tells us anything important about minds may also want our explainer on whether passing the Turing test means AI is conscious.

A narrower reading is also available from the same text. The guidance says the AI Act does not apply in Northern Ireland in full, and it repeatedly limits coverage to AI closely integrated with physical products placed on the market. The Commission’s Article 50 text also keeps an exception where AI use is obvious from the context. So the answer to does AI have to say it is AI is not a general yes. In the cases the guidance describes, it is a conditional yes for some product providers, and a no for many standalone tools.

The guidance also cautions against treating every in-scope product as heavily regulated. It states: “Some AI systems may not face any specific requirements under the EU AI Act.” For broader context on where disclosure disputes sit in the field, see the timeline of the AI rights debate.

Frequently asked questions

Does every AI system in Northern Ireland have to identify itself?

No. The 4 November 2026 guidance says the rule applies only where an AI system is closely integrated with a physical product placed on the Northern Ireland market. The same guidance says standalone software, web tools, apps and APIs are generally out of scope unless they are closely integrated with a physical product.

Do online chatbots have to disclose that they are AI in Northern Ireland?

Not under the Northern Ireland scope described in the guidance, if they are just web-based or downloadable tools. The guidance gives a generative AI chatbot available for download or via the web as out of scope. It also says that if a laptop manufacturer installs a generative AI chatbot on a laptop and places that laptop on the Northern Ireland market, the chatbot becomes in scope because it is integrated with a physical product.

Is the disclosure rule only about telling users they are talking to AI?

No. The guidance says Article 50 has two relevant parts in Northern Ireland. Article 50(1) covers informing users that they are interacting with AI, from 10 November 2026. Article 50(2) covers marking AI-generated synthetic audio, image, video or text content so it is machine-readable and detectable, from 2 December 2026.

When do the higher-risk Northern Ireland rules start?

The guidance says Annex III high-risk obligations apply from 2 December 2027, and Annex I high-risk obligations apply from 2 August 2028. Those dates are later than the 10 November 2026 start date the guidance gives for Article 50(1) disclosure duties.

Caveats: The Northern Ireland-specific scope and dates in this briefing come from the 4 November 2026 guidance document. The European Commission’s Article 50 text sets out the underlying transparency rule, but it does not itself state Northern Ireland’s 10 November 2026 start date or the Windsor Framework limits described here.

Sources

  1. UK Government, Guidance on the EU AI Act in Northern Ireland — 2026-11-04
  2. European Commission AI Act Service Desk, Article 50: Transparency obligations for providers and deployers of certain AI

Tags: northern-ireland · eu-ai-act · ai-disclosure · transparency

← all briefings