Law, policy and the UK

Online Safety Act

In short

The Online Safety Act is the UK law, passed in 2023, that makes social media, search engines and other online services protect users from illegal content and children from harmful content. Ofcom enforces it.

Also called
Online Safety Act 2023, OSA
Topic
Law, policy and the UK
Level
Beginner
First described
Passed by Parliament and given Royal Assent on 26 October 2023 as the Online Safety Act 2023 (c. 50)1.

The Online Safety Act 2023 received Royal Assent on 26 October 2023. Its long title describes it as an Act to provide for "the regulation by OFCOM of certain internet services"1.

It mainly covers two kinds of service: user-to-user services, where people can share content with each other, and search services. They must assess the risks on their platform and act on them. The duties to tackle illegal content became enforceable on 17 March 2025, and the children's safety rules came fully into effect in summer 20252. Sites that publish pornography have had to use age checks since 17 January 20252. Ofcom can fine a company up to £18 million or 10% of its qualifying worldwide revenue, whichever is greater2.

AI is covered where it fits those categories. In November 2024 Ofcom told providers that a chatbot whose answers users can share with others is a user-to-user service, that chatbots made by users count as user-generated content, and that AI tools which search across websites are search services3. AI tools that can produce pornography must use "highly effective age assurance"3.

That left a gap. A standalone chatbot that only talks to its own user may fall outside the Act, and Ofcom said chatbot-made images are generally out of scope unless they are pornographic or shareable4. In February 2026 the government said it would close the gap through the Crime and Policing Bill4. The resulting Crime and Policing Act 2026, which received Royal Assent on 29 April 20265, creates an offence for making available AI chatbots that generate certain prohibited content and amends section 216 of the Online Safety Act so ministers can make regulations on illegal AI-generated content6.

Example

If a UK social network adds an AI chatbot and lets users post its replies into group chats, Ofcom treats that feature as part of a user-to-user service, so the platform's illegal content duties apply to it3.

Why it matters

It is the main UK law deciding what online platforms, including some AI chatbots, must do to keep users safe. It explains many of the age checks and content controls UK users now see.

Questions people ask

Does the Online Safety Act apply to ChatGPT and other AI chatbots?

Partly. Ofcom says chatbots are covered when users can share their outputs, when users build chatbots for others, or when the tool searches the web3. Standalone chatbots were a gap4, which the Crime and Policing Act 2026 moves to close by letting ministers make regulations on illegal AI-generated content6.

Who enforces the Online Safety Act?

Ofcom, the UK communications regulator. It can fine companies up to £18 million or 10% of qualifying worldwide revenue, whichever is greater2.

When did the Online Safety Act come into force?

It became law on 26 October 20231 but applies in stages. Age checks for porn sites applied from 17 January 2025, illegal content duties became enforceable on 17 March 2025, and children's safety rules took full effect in summer 20252.

Related terms

Sources

  1. Online Safety Act 2023 (c. 50), introductory text, legislation.gov.uk
  2. DSIT, Online Safety Act: explainer, GOV.UK
  3. Ofcom (8 November 2024), Open letter to UK online service providers regarding generative AI and chatbots
  4. Lewis Silkin (2026), Online safety reforms to be fast tracked amid rising AI risks
  5. Home Office, Crime and Policing Act 2026 collection, GOV.UK
  6. Clyde & Co (2026), The Crime and Policing Act 2026: intimate image abuse and AI risks

Back to the glossary