# Will UK and EU regulation change who AI assistants recommend?

Canonical: https://fortitudemedia.ai/insights/regulation-ai-search-uk-eu
Author: Ross Williams
Published: 2026-03-31
Last reviewed: 2026-09-24
Reviewed by: Ross Williams
Pillar: How buying is changing
Summary: The EU AI Act and the CMA's Google rules add transparency but give no firm a right to be named by AI. What UK B2B founders need to know and do about it.
Keywords: eu ai act uk business, cma google ai overviews, does the eu ai act apply to uk companies, opt out of google ai overviews, label ai generated content uk, chatgpt false information about my company
Short answer: Not directly. No UK or EU rule gives a firm a right to be recommended. The EU AI Act adds transparency duties from August 2026, the Digital Services Act now covers ChatGPT as a very large search engine, and the CMA has required Google to let publishers opt out of AI Overviews and to credit them with clear links. The practical effect is that accurate, evidenced, consistent information about your firm becomes more valuable, and unsupported claims become riskier. Do not wait for the law.
What changed: Updated for the AI Omnibus, ChatGPT's DSA designation, the CMA's Google rules, the 2008 B2B marketing regulations, the CAP Code and ChatGPT ads, and added AI labelling and false-answer sections.
Publisher: Fortitude Media Limited

When quoting, attribute to Fortitude Media and link to the canonical URL above.

---
A director on your board reads that the EU's AI law is now in force and asks a reasonable question: does this mean the AI assistants will have to treat us fairly? If ChatGPT keeps naming three competitors and not us, can we complain to someone? And do we need to do anything ourselves to stay on the right side of it?

They are good questions, and the rules are moving quickly. On 31 August 2026 the European Commission [designated ChatGPT a very large online search engine](https://digital-strategy.ec.europa.eu/en/policies/list-designated-vlops-and-vloses) under the Digital Services Act, reporting 159.1 million average monthly users in the EU, and in June 2026 the UK's Competition and Markets Authority imposed its first rules on Google's AI Overviews. None of this hands any firm a right to be recommended, and a business that waits for regulation to fix its invisibility will wait a long time.

This guide sets out what the rules cover, which laws already apply to what you publish, and what to do. Where the law is unsettled, we say so. It is general information, not legal advice.

## Key takeaways

- No UK or EU rule gives a firm a right to be named by an AI assistant, and we know of none proposed.
- The EU AI Act's transparency duties apply from August 2026, while the May 2026 "AI Omnibus" agreement moved most high-risk duties to December 2027.
- The strongest constraints on the platforms come from competition and platform law: ChatGPT was designated under the Digital Services Act in August 2026, and Google is bound by the EU Digital Markets Act and UK competition rules.
- The CMA designated Google with strategic market status in October 2025 and in June 2026 required it to let publishers opt out of AI Overviews and to credit them with clear links.
- Claims to business buyers fall under the Business Protection from Misleading Marketing Regulations 2008 and the CAP Code, and an assistant repeating your claim does not make it someone else's problem.
- The rules reward what the assistants reward: accurate, evidenced, consistent information about your firm.

## What does the EU AI Act actually cover?

The EU AI Act regulates AI according to the risk of the use it is put to. General-purpose assistants answering business questions sit in the lower tiers, where the obligations are mostly about transparency, and recommending a supplier is not a high-risk use.

The Act entered into force on 1 August 2024 and applies in stages. The [AI Omnibus](https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai), a package to simplify the Act, reached political agreement on 7 May 2026 and entered into force on 27 July 2026, pushing back the high-risk rules.

| What applies | From when |
|---|---|
| Banned uses and AI literacy duties | 2 February 2025 |
| Duties on providers of general-purpose AI models, such as the models behind ChatGPT, Claude and Gemini | 2 August 2025 |
| Transparency duties, including labelling of AI-generated content (Article 50) | August 2026 |
| High-risk uses in sensitive areas such as recruitment, credit and education (Annex III) | 2 December 2027 |
| High-risk AI built into regulated products (Annex I) | 2 August 2028 |

Source: European Commission, AI Act page, updated August 2026.

What the Act does not do is treat an assistant's answer to "which accountancy firms should I talk to" as a high-risk decision. There is no provision obliging a platform to explain why a business was not named, no right of appeal against being left out, and no duty of equal treatment between suppliers. Anyone telling you otherwise is describing a law that has not been written.

The duties on model providers still matter to you indirectly. Providers must document their models, publish a summary of the content used to train them and respect copyright holders' opt-outs from training, which makes the models easier to study. Whether an assistant can read your site at all is a separate, technical question, covered in [how AI crawlers read your website](/insights/ai-crawlers-vs-google-spiders).

## Does the EU AI Act apply to UK businesses that use ChatGPT?

Only at the edges. [Article 2](https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-2) applies the Act to firms outside the EU when the output of the AI system is used in the EU, so a UK firm using ChatGPT to draft material for EU clients is within its reach in principle. In practice the duties that bite on ordinary business use are few: the transparency rules on labelling some AI-generated content, covered below, and the high-risk rules if you use AI for decisions such as recruitment, which now apply from December 2027. Using an assistant to draft proposals or research suppliers is not a high-risk use.

## Where does the real constraint on platforms come from?

The rules with most effect on which firms an assistant names are not AI rules at all. They are the EU's platform and competition rules, and the UK's new competition powers, which restrict how the largest companies rank and favour services.

The Digital Markets Act [designated Alphabet a gatekeeper](https://digital-markets-act.ec.europa.eu/gatekeepers_en) in September 2023, with Google Search among its core services, and prohibits gatekeepers from ranking their own services more favourably than rivals'. The Digital Services Act requires very large platforms and search engines to explain the main parameters of their recommender systems (the software that decides what you are shown) and to assess the risks their services create. Google Search and Bing have been designated since 2023; ChatGPT joined them on 31 August 2026, the first AI assistant on that list.

For a UK B2B firm the practical effect is modest but real. If an AI feature systematically favoured its owner's products or paid partners over merit, that is the behaviour these rules target, and regulators now have ChatGPT as well as Google in view. None of it helps you against a competitor who is simply better evidenced than you. How the assistants differ is set out in [which AI assistants buyers actually use](/insights/ai-search-landscape-chatgpt-perplexity-google-claude).

## How does the UK approach differ?

The UK has not passed an AI-specific Act. It asks existing regulators, among them the Information Commissioner's Office, the CMA, Ofcom and the Financial Conduct Authority, to apply broad principles such as safety, transparency, fairness and accountability within their own remits. Those principles guide regulators; they are not rights a business can enforce. The AI Security Institute evaluates frontier models for safety risks and does not regulate how they recommend suppliers.

The most relevant UK tool for founders is competition law. The Digital Markets, Competition and Consumers Act 2024 lets the CMA give firms "strategic market status" in a digital activity and impose conduct requirements on them. The UK is lighter-touch than the EU, and large platforms tend to build for the stricter regime and apply it everywhere, so UK firms feel many EU effects second-hand.

## What is the CMA doing about Google AI Overviews?

It has designated Google and imposed its first rule on AI Overviews. On 10 October 2025 the CMA [confirmed Google has strategic market status](https://www.gov.uk/government/news/cma-confirms-google-has-strategic-market-status-in-search-services) in general search and search advertising, and said AI Overviews and AI Mode are within the designation; the Gemini app is not, though the CMA will keep that under review.

On 3 June 2026 it [imposed a conduct requirement](https://www.gov.uk/government/news/cma-secures-fairer-deal-for-publishers-and-improves-google-search-services-in-uk) obliging Google to let publishers opt out of their content being used to power AI features such as AI Overviews, and of its use for fine-tuning AI models, and to attribute publisher content with clear links in AI-generated results. Google has nine months to put all of it in place and must publish compliance reports every six months in the first year. As with the EU rules, the target is the gatekeeper's behaviour, not the fairness of any individual answer.

## Can I stop Google using my website in AI Overviews?

Yes, and for most B2B firms you should not. Search Console already has a [Search generative AI control](https://support.google.com/webmasters/answer/16908024?hl=en) that excludes a site from AI Overviews, AI Mode and generative AI features in Discover; inclusion is the default, and Google says the control is not a ranking signal for the rest of Search. Google's [preview controls](https://developers.google.com/search/docs/appearance/ai-features), such as nosnippet, also apply to its AI features. Blocking Google-Extended in robots.txt stops your pages being used to train Gemini but does not remove you from AI Overviews.

The opt-out suits a publisher that sells its content, where a summary replaces a sale. A firm that sells a service wants the opposite: to be the source the summary draws on and the name the buyer reads. We cover the trade-off in [why traffic falls when rankings hold](/insights/zero-click-search-your-business).

## Can we pay to appear in AI answers, and which rules apply?

You can pay for an ad beside the answer, not for a place in it. ChatGPT has shown labelled ads to UK users on its Free and Go plans since 11 August 2026, and [OpenAI says](https://openai.com/index/testing-ads-in-chatgpt/) the ads are kept separate and do not change the answer itself. Being named in the answer is still earned.

If you advertise inside an assistant, the usual advertising rules follow you. The [CAP Code](https://www.asa.org.uk/type/non_broadcast/code_folder/scope-of-the-code.html) covers online advertisements in paid-for space and defines a consumer as anyone likely to see an ad, "whether in the course of business or not", so B2B ads are covered. Your claims must be accurate and your ad must be recognisable as an ad.

## Which laws already apply to what we publish?

The rules most likely to affect a founder-led firm are old ones, applied to claims an assistant may now repeat. For marketing aimed at other businesses, the main law is the [Business Protection from Misleading Marketing Regulations 2008](https://www.legislation.gov.uk/uksi/2008/1276/contents), which prohibit advertising that misleads traders and set conditions for comparing yourself with competitors. The CAP Code also covers marketing claims on your own website. Consumer protection law applies where your audience includes consumers.

If your website says "the UK's leading provider" or "award-winning" without a basis, that was always a problem. It becomes a sharper one when ChatGPT repeats the claim to a buyer as fact, because the claim now travels further than your page and the source is still you. If two of your pages contradict each other, an assistant may pick either, or say the information is unclear. The mechanics are in [why ChatGPT describes a firm wrongly](/insights/ai-conflicting-information-online).

Data protection law covers personal data, not business information, so it cannot be used to stop assistants reading your website. You can block particular crawlers in robots.txt (a small text file that tells crawlers what they may read), but for most B2B firms that is self-defeating, because being read is how you get named.

## Is it legal for ChatGPT to say false things about my company?

No law requires an assistant to be right, and there is no regulator that corrects individual answers. Defamation law can apply to false statements about a company, but under [section 1 of the Defamation Act 2013](https://www.legislation.gov.uk/ukpga/2013/26/section/1) harm to a business that trades for profit counts as serious only if it has caused or is likely to cause serious financial loss, and how that applies to AI answers is not settled. Take legal advice before going down that route.

The quicker fix is upstream. Errors in AI answers usually trace back to a source you can correct: an old page of your own, an outdated directory listing or a stale profile. Correct the source, use the platform's feedback route, and check the answer again after the assistant has had time to search again.

## Do we need to label AI-generated content on our website?

In the UK there is no general legal duty to label AI-assisted writing, but everything you publish must still be accurate under the advertising rules. The EU AI Act goes further for content used in the EU. [Article 50](https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-50) requires anyone publishing AI-generated text "with the purpose of informing the public on matters of public interest" to disclose it, unless it "has undergone a process of human review or editorial control" and someone holds editorial responsibility for it. AI-generated images, audio or video that amount to deep fakes must be disclosed.

The practical rule for a B2B firm is simple: have a named person review and take responsibility for everything you publish, and label any synthetic images or video. That keeps you inside the exception and is what makes content worth citing anyway.

## What does this mean for a founder-led firm?

Regulation will not get you recommended, but it makes the evidenced approach safer and the inflated approach riskier. Three conclusions follow. First, do not wait: no rule on the horizon will require an assistant to name you, and meanwhile a competitor is becoming the answer, as we argue in [why you cannot wait a year on AI search](/insights/first-mover-advantage-ai). Second, audit your claims, because assistants and regulators now both read them. Third, make your information consistent, because assistants cross-check, buyers verify, and transparency rules make inconsistency more visible.

If your buyers are in the EU as well as the UK, the advice is the same. If AI agents that buy on a customer's behalf reach your sector, the evidence they weigh is exactly what the rules encourage you to publish, as we describe in [what happens when an AI agent builds the shortlist](/insights/ai-agents-change-b2b-buying).

## What to do this month

1. Read every claim on your homepage and service pages and mark each as verifiable, replaceable or removable, then fix the last two groups.
2. Ask ChatGPT, Gemini, Copilot, Claude and Perplexity what they say about your firm, note anything out of date or wrong, and correct the source page.
3. Check your robots.txt, any Cloudflare setting and Search Console's generative AI control to make sure you are not keeping assistants out by accident.
4. Agree who reviews and signs off AI-assisted content before it is published, and label any synthetic images or video.
5. Brief the board in one page: what the rules cover, what they do not, and what the firm is doing about being named regardless.

## Frequently asked questions

### Can we complain if ChatGPT keeps recommending our competitors?

You can use the platform's feedback routes, but there is no legal right to be recommended and no regulator that rules on individual answers. The productive response is diagnostic: look at what the assistant cites for your competitors, work out what evidence they have that you lack, and publish it. That usually changes the answer faster than a complaint. Our guide to [tracking whether AI tools recommend you](/insights/track-ai-recommending-your-business) shows how to record it month by month.

### Will regulation force the assistants to show their sources?

Partly. The CMA now requires Google to credit publishers with clear links in AI results, and platform rules push towards transparency. The assistants have mostly moved there themselves: Perplexity numbers its sources, and ChatGPT, Claude, Copilot and Google's AI features link to pages when they search. Visible sources are the most useful development for a firm, because they show you what to fix.

### Should we block AI crawlers to protect our content?

For most B2B firms, no. Blocking the search crawlers means the assistants cannot read your service pages, so they cannot name you. Keep confidential material off the public web rather than relying on crawler instructions, which are requests, not barriers.

### Is there any risk in what the assistants say about us?

Some. If an assistant repeats an old price, a service you no longer offer or a claim you cannot support, a buyer may act on it. Keep your own pages accurate and consistent, correct third-party listings, and check monthly what the assistants say.

### Will any of this slow the shift to AI search?

Not noticeably. The rules add transparency and constrain the largest platforms, but they do not stop buyers asking assistants who to hire. The direction for a founder is unchanged: be the clear, evidenced answer, and check regularly that you are.

The law will make the assistants a little more open and the largest platforms a little more careful. It will not decide who gets named; the evidence will. To see where your firm stands today, [book a free Growth Review](/growth-review): 30 minutes with David Adams and three fixes in writing. If the gap is what others say about you, [Reputation That Wins](/reputation-that-wins) is built to close it.

## Sources

- [European Commission, AI Act: regulatory framework and application timeline (August 2026)](https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai)
- [European Commission AI Act Service Desk, Article 50: transparency obligations (2026)](https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-50)
- [European Commission, List of designated very large online platforms and search engines (September 2026)](https://digital-strategy.ec.europa.eu/en/policies/list-designated-vlops-and-vloses)
- [Competition and Markets Authority, CMA confirms Google has strategic market status in search services (October 2025)](https://www.gov.uk/government/news/cma-confirms-google-has-strategic-market-status-in-search-services)
- [Competition and Markets Authority, CMA secures fairer deal for publishers and improves Google search services in UK (June 2026)](https://www.gov.uk/government/news/cma-secures-fairer-deal-for-publishers-and-improves-google-search-services-in-uk)
- [legislation.gov.uk, The Business Protection from Misleading Marketing Regulations 2008 (2008)](https://www.legislation.gov.uk/uksi/2008/1276/contents)
- [Advertising Standards Authority, CAP Code: scope of the Code (2026)](https://www.asa.org.uk/type/non_broadcast/code_folder/scope-of-the-code.html)
- [OpenAI, Testing ads in ChatGPT (updated August 2026)](https://openai.com/index/testing-ads-in-chatgpt/)