The UK Online Safety Act and Uncensored AI: Age Checks Explained

Whether the UK Online Safety Act reaches an AI chat depends on what the chatbot does, not on what it calls itself. For UK Online Safety Act age verification, Ofcom's test has three parts. A chatbot is regulated if users can share its output with each other, if it searches more than one website or database, or if it can generate pornographic material. A chatbot that meets none of the three is not regulated at all, though the government has said it will look again.

The rule that already bites is the pornography one. Any service that allows pornography and has links to the UK needs strong age checks, and Ofcom says ticking an "I am over 18" box does not count. Plans to go further, such as limits for under-18s on sexualised chatbots, are announced but not law.

Everything below comes from Ofcom, gov.uk, legislation.gov.uk and ICO pages read on 13 October 2026. We did not test any platform's age checks, and we did not consult a lawyer, so this is general information, not legal advice. HackAIGC publishes this post, and it makes no claim that HackAIGC or any other service meets the UK rules.

What Does the UK Online Safety Act Mean for AI Chatbots?

The Act does not mention chatbots by name. Ofcom applies its existing categories to them, so whether a chatbot is regulated depends on how it works, not on whether it is called AI. The Act regulates user-to-user services, search services, and services that publish pornography.

Ofcom set this out in an open letter to online service providers dated 8 November 2024. It says a site or app with a generative AI chatbot is a user-to-user service where users can share the chatbot's output with other users. It says the same applies where users can upload or create their own chatbots. It says "Generative AI tools that enable the search of more than one website and/or database" are search services. And it says sites and apps with tools that can generate pornographic material are also regulated.

In February 2026 Ofcom went further. In a scope update published 3 February 2026, it wrote: Chatbots are not subject to regulation at all if they" meet three conditions. They only let people interact with the chatbot itself and no other users. They do not search multiple websites or databases. And they cannot generate pornographic content.

That leaves a gap, and Ofcom said so. The same update says "The Secretary of State has said that the Government will look at how chatbots should be regulated."

The three conditions work like a checklist. A private chat app where only you see the answers, with no web search and no adult image or text output, sits outside the Act. Add a shared feed, a web search tool, or the ability to produce adult material, and the same app moves inside it. One feature change can alter a service's legal position, which is why Ofcom frames the question around functions.

For uncensored AI the pornography condition matters most. A platform that can produce adult images or text is in scope for Ofcom's pornography duties even when no other rule applies. HackAIGC's guide to whether NSFW AI is legal covers how different countries treat adult AI content.

What Counts as a Valid Age Check Under Ofcom's Rules?

For UK Online Safety Act age verification, Ofcom accepts age checks that are "highly effective" and rejects self-declaration. Its age checks page says its guidance lists methods "capable of being highly effective." That list is not exhaustive.

The methods Ofcom names are open banking, photo ID matching, facial age estimation, mobile network operator age checks, credit card checks, digital identity services and email-based age estimation.

The same page says what does not count. It confirms that methods including self-declaration of age, and online payments which don't require a person to be 18, are not highly effective. In practice, a box that asks "Are you 18?" does not meet the standard. A payment alone does not either, unless the payment method itself confirms the buyer is an adult.

Ofcom's letter to parliamentarians on 16 January 2025 gives the test behind the list. It says age assurance must be technically accurate, robust, reliable and fair.

As for this site, HackAIGC's terms of service say users must be at least 18. Its privacy policy says "Our services are only available to users who are 18 years or older." Neither document names a method, such as ID matching or facial age estimation. Under Ofcom's description, a self-declared rule on its own is not highly effective.

This post makes no claim that HackAIGC meets the UK rules, and you should not read one into it.

Strong checks have a cost for users as well as providers. Every method on Ofcom's list processes personal data. That is why the privacy terms of the check matter as much as the method. The joint Ofcom and ICO statement of 25 March 2026 says: "We recognise that all age assurance methods involve the processing of personal data." It also says "self-declaration alone is not an effective means to determine the age or age range of users."

Which Parts of the Law Are Already in Force, and Which Are Still Proposed?

Ofcom's pornography duty has been in force since January 2025, and the strong age checks it requires have been enforced since July 2025. The power to extend the Act to all AI chatbots has existed since April 2026. The government's own July 2026 response says the detailed chatbot definitions are still to come in regulations. The table lists the dated steps.

Sources: the gov.uk Online Safety Act explainer for the January 2025 duty, the Ofcom pages cited above for the guidance rows, and legislation.gov.uk for the statutes.

Section 248 is titled "Power to amend Online Safety Act 2023: AI." It inserts a new section 216A that lets the Secretary of State amend the Act by regulations to reduce harm from illegal AI-generated content. The legislation page states that section 248 is in force at Royal Assent.

A power is not the same as a rule. Section 249 requires the Secretary of State to lay a progress report "no later than 31 December 2026," unless a draft statutory instrument under that power is laid first. Until regulations are made, the chatbot duties the government has described are proposals.

The government's July 2026 response, on gov.uk, says it will "ensure that all chatbots are required to take action on illegal content." It also describes plans to "prevent under-18s from accessing AI chatbot services that primarily offer sexualised content." It adds that "the majority of chatbots most used by children are already within scope of the Online Safety Act," and that detailed definitions "will be set out in regulations in due course." None of that has a commencement date on the page, so treat it as a plan. The same response rules out the VPN restriction floated in February, which the FAQ below covers.

The same Act created a separate offence in section 99: making, adapting, supplying or offering to supply a thing "for use as a generator of purported intimate images." Section 99 came into force on 29 June 2026 and carries up to 3 years in prison on conviction on indictment.

How Is Ofcom Enforcing the Rules in 2026?

Ofcom has fined pornography services for failing to use age assurance, and it says its duties apply to any adult service with links to the UK. Its enforcement programme page says: "The duties apply to all adult services that have 'links to the UK', regardless of where the provider is based."

The latest penalties for age-assurance failures were £700,000 for xgroovy.com on 3 September 2026 and £600,000 for fapello.com on 8 July 2026. Youngtek Solutions Ltd was fined £500,000 on 26 May 2026.

Earlier ones on the same page include £1,350,000 for 8579 LLC in February 2026, £800,000 for Kick Online Entertainment S.A. in February 2026, and £1,000,000 for AVS Group Ltd in December 2025. Several of these also carried a smaller penalty for failing to answer an information request.

The legal ceiling is higher. The gov.uk explainer says "Companies can be fined up to £18 million or 10 percent of their qualifying worldwide revenue, whichever is greater." It also says Ofcom can ask a court to make payment providers, advertisers and internet service providers stop working with a site.

Ofcom has also looked at the AI companion market directly. On 15 January 2026 it opened an investigation into Novi Ltd, which runs an AI character companion chatbot service, to check its compliance with age check requirements under the Act. Ofcom's page says sites that allow pornographic material "must use highly effective age assurance to prevent children from readily accessing that content." The investigation was ongoing when we read it, and it has reached no finding.

The Grok case shows the limits of Ofcom's reach. Ofcom's scope update says "we are currently unable to investigate the creation of illegal images by the standalone Grok service in this case." It adds: "We can only take action on online harms covered by the Act, using the powers we have been granted." Ofcom opened its separate investigation into X over Grok on 12 January.

The page lists eight penalties between November 2025 and September 2026. All of them landed on adult content services. The chatbot investigation into Novi Ltd has no finding yet, so treat today's record as a snapshot and not a forecast.

Why Do Some AI Platforms Restrict or Block UK Users?

Several providers on Ofcom's own enforcement page have blocked UK users. Itai Tech Ltd, Duplanto Ltd, First Time Videos LLC and Xgroovy are each recorded as having implemented a block on UK IP addresses. The page records Itai Tech's investigation as closed on 15 January 2026 and Duplanto's on 30 March 2026.

The rules create three pressures that bear on a decision like that. First, jurisdiction: Ofcom says the duties follow the UK link, not the company's location, so a small overseas provider is covered. Second, the fine ceiling of £18 million or 10 percent of qualifying worldwide revenue is large for any small service. Third, every strong method processes personal data. The joint Ofcom and ICO statement says all age assurance methods involve processing personal data, so a platform that adds ID checks takes on privacy duties too.

For users the effect is uneven. One platform may add an ID or face check, another may block UK visitors, and a third may add no new checks. You cannot tell which from the outside, so judge each service on what it publishes.

HackAIGC's NSFW AI privacy guide explains how to think about data when you use any adult AI tool.

How Can You Judge an AI Platform's Age Checks and Privacy?

Ask four questions: how does it check age, what data does that check collect, who holds the data, and what does its policy say about deletion. A platform that answers all four in plain language has thought about the problem. One that answers none has not.

Start with the method. Compare what the platform describes with Ofcom's list: open banking, photo ID matching, facial age estimation, mobile operator checks, credit card checks, digital identity or email-based estimation. A bare "I am over 18" button matches Ofcom's example of a method that is not effective.

Then look at the data. The joint Ofcom and ICO statement says all methods involve personal data. A good policy says what is collected, for how long, and who the third-party provider is.

Next, look for honesty about limits. A platform that says "we use self-declaration" is telling you the truth even if the method is weak. One that implies compliance without describing a method deserves more doubt.

Finally, check what the law says about content, because age checks are only one part. HackAIGC's guide to deepfake face and body swap tools covers consent and legal risk, and its piece on AI content censorship covers how rules differ by region. If you wonder why mainstream assistants refuse adult requests at all, see why ChatGPT will not generate NSFW. And for the basics of the category, read what uncensored AI is.

Frequently Asked Questions

Is There a UK Law Regarding the Use of AI?

The UK has no single law that covers AI as a technology. A House of Commons Library briefing published on 10 June 2026 says: "The UK does not have any AI-specific regulation or legislation covering AI as a technology." Narrow laws do apply: the Online Safety Act, the offence of creating a purported intimate image without consent, and section 248 of the Crime and Policing Act 2026.

Is ChatGPT Banned in the UK?

No. OpenAI's help page lists the United Kingdom among the countries where ChatGPT is supported on web and mobile. Whether any chatbot falls under Ofcom's rules depends on how it works: user sharing, search, or pornographic content. A chatbot that does none of those three is not regulated under the Act, according to Ofcom's February 2026 update.

What Is Illegal to View on the Internet in the UK?

Illegal content under the Act includes extreme pornography and intimate image abuse. The gov.uk explainer says illegal content includes "harassment, stalking, controlling or coercive behaviour, extreme pornography, and intimate image abuse." Possession of an extreme pornographic image is an offence under section 63 of the Criminal Justice and Immigration Act 2008. Ask a solicitor if you are unsure about a specific case.

Will the UK Ban VPNs for Children?

Not at present. The government's July 2026 response says "We will not ban or age-gate VPNs." That is a change from its February 2026 announcement, which said the consultation would look at options to "age restrict or limit children's VPN use where it undermines safety protections." The response says Ofcom will research how services can detect and reduce circumvention, including through VPNs, and asks VPN providers to work with government on safer products. It also says: "We are prepared to take further action on this, should evidence show that circumvention is not being tackled adequately."

Is the UK Censoring Free Speech?

Section 22(2) of the Act requires providers, when deciding on and implementing safety measures, to have "particular regard to the importance of protecting users' right to freedom of expression within the law." Ofcom's scope update also says it can only take action on online harms covered by the Act.

Is It Illegal to Create AI-Generated Intimate Images of Real People in the UK?

Creating a purported intimate image of another person without consent became an offence on 6 February 2026, under section 138 of the Data (Use and Access) Act 2025. Sharing a photograph or film that shows, or appears to show, a person in an intimate state without consent is covered by section 66B of the Sexual Offences Act 2003, inserted by the Online Safety Act and in force since 31 January 2024. Making or supplying a generator of such images became an offence on 29 June 2026.

Can Ofcom Fine a Company Based Outside the UK?

Ofcom says its duties apply to all adult services that have links to the UK, regardless of where the provider is based. The maximum fine is the greater of £18 million or 10 percent of qualifying worldwide revenue. The gov.uk explainer also says Ofcom can ask a court to make payment providers, advertisers and internet service providers stop working with a site.

Do Age Checks Apply to Text-Only AI Chat?

Not by themselves. Ofcom says a chatbot is unregulated if it only lets users interact with the chatbot, does not search several sites or databases, and cannot generate pornographic content. If text chat can produce pornographic material, or users can share outputs, the picture changes. The government has also said it will look at how chatbots should be regulated.

Are Ofcom's Rules Only About Pornography?

No. The Act also covers illegal content such as harassment, stalking and intimate image abuse, and it adds duties to protect children. Pornography is where the strong age-check requirement is clearest, and where Ofcom has fined companies so far. Other duties apply to user-to-user and search services, and which ones depend on how a service works.

Related Articles

• AI Porn Generators: What Is Legal and What Is Not 2026

• NSFW AI Ethics Debate: Uncensored AI Is Here, What Now? (2026)

• Is NSFW AI Safe? Privacy, Security & Legal Guide (2026)

• The Rise of Uncensored AI: Why Everyone Is Switching in 2026

Try Uncensored Chat on HackAIGC

HackAIGC's free plan gives you 6 requests a day, and paid plans are one-time payments with no auto-renewal. You must be 18 or older. As noted above, this post does not claim the site meets UK age-check rules.

Try HackAIGC free today →