The UK Online Safety Act

OpenAI is committed to complying with the UK Online Safety Act⁠⁠(opens in a new window). The UK Online Safety Act imposes duties on providers of services regulated by the Act relating to illegal content and content harmful to children. We work hard to promote responsible use of our products and keep our users safe.

We aim to balance delivering helpful and accessible information to all users while mitigating the risks of online harm. We use a range of procedures and tools to protect users from illegal and harmful content. 

This includes measures to prevent, detect, respond to, and take enforcement action against illegal content, such as terrorism content, child sexual exploitation and abuse content and other illegal content. 

We aim to review and remove illegal content as swiftly as possible when we become aware of it, whether via our own proactive detection methods, or from reports from third parties, including our users. This helps to prevent users encountering such content and minimises the length of time illegal content is present on the service. 

More information on our moderation and enforcement processes is set out in our Transparency & Content Moderation page⁠. Please see our Reporting content page⁠⁠(opens in a new window), for details about how you can report content, including illegal content, on our services.

We aim to provide a safe online experience for all our users, and take action to protect all users (including users under 18) from harmful content that violates our policies. This includes content the UK Online Safety Act⁠⁠(opens in a new window) recognises as content harmful to children*.

When we become aware of illegal or harmful content, we take appropriate action, balancing the importance of protecting our users and ensuring they have access to information. For example: 

• On ChatGPT Search: We take action to protect users from encountering illegal or harmful content via search results, while aiming to ensure our users have access to the information most relevant to their search query.

If you think you have encountered illegal or harmful content on our services, please report it to us⁠⁠(opens in a new window) and we will investigate. We are committed to user safety and will take action to help prevent violative content from being provided to other users. 

There are also third party resources recommended by Ofcom⁠(opens in a new window) available to you, if you or someone you know has experienced serious harm online.

We use proactive technology to help prevent users encountering illegal content and harmful content on our services. This includes content classifiers, reasoning models, hash-matching and other automated systems to identify content that may violate our terms or policies. 

More information, including about our moderation and enforcement processes is set out in our Transparency & Content Moderation page⁠.

We use age prediction technology to provide age-appropriate experiences on our services. This can include extra safeguards for sensitive or age-inappropriate content and some feature limits. For more information on age prediction, please visit our Age prediction page⁠(opens in a new window).

If you’re in the UK and think OpenAI isn't complying with its obligations under the UK Online Safety Act or has used proactive technology to moderate content in a way that is not compliant with our terms⁠, you can report this to us via our UK Online Safety Act Reporting Form⁠.

We will review your report and consider how your feedback may help us improve our processes. 

We aim to complete reviews of reports within 10 business days although this process may take longer for more complex reports. We’ll follow up with you only if we need more information or have additional information to share with you.

If we take enforcement action based on your content or activity (including following our use of proactive technology), and you think we have a mistake, you can report this to us and appeal our decision. Further information on how to appeal is set out on our Transparency & Content Moderation page⁠. 

We aim to review appeals promptly, though more complex cases may take longer. 

Trusted flaggers are designated organizations that work with us either by agreement or as may be required by law. Material reported by trusted flaggers is reviewed against our terms and policies, just like all other reports, and then actioned accordingly. 

Trusted Flaggers can submit reports using their organization email address through either the Report Content Form or the UK Online Safety Act Reporting Form⁠.

* Under the UK Online Safety Act content harmful to children includes pornography, content which encourages or promotes suicide, self-harm or eating disorders, hate speech, bullying content, content which depicts realistic graphic violence, content which encourages or promotes serious violence against a person, dangerous challenges likely to result in serious injury, content that actively encourages taking physically harmful drugs or substances, or content that promotes body stigma and depression.