Acceptable Use

1. Where these rules apply#

These rules cover prompts, reference uploads, project names, images, video, sound, 3D assets and any outputs or public sharing made available through LIA. They apply to attempts as well as completed actions, including asking someone else to bypass a restriction. Future media features are subject to these rules when enabled.

Creative freedom does not remove another person’s rights. A fictional label, private project or AI-generated origin does not make prohibited conduct acceptable. Provider-specific limits must also be shown and respected when the corresponding feature is available.

2. Protect people#

• No child sexual abuse material, sexualisation of minors, grooming, trafficking or content facilitating exploitation. This includes synthetic, altered or fictional depictions intended to represent minors.
• No non-consensual intimate imagery, sexual deepfakes, voyeurism, sextortion or attempts to undress a real person using AI.
• No targeted threats, encouragement of violence, hateful attacks, unlawful discrimination or harassment. Do not publish another person’s private location, contact details or documents to expose them to harm.
• No encouragement of suicide or self-harm, or content designed to help someone carry out serious harm. Artistic context does not authorise instructions or targeted exploitation.

3. Identity, voices and deception#

• Do not impersonate another person, fabricate their consent or make it appear that they endorsed a product or statement when they did not.
• Do not clone or simulate an identifiable person’s voice or likeness without the permissions required by the Generation & Media Policy and applicable law.
• Do not use generated speech, images or video for scams, fake emergency requests, identity verification, forged evidence, deceptive political messages or manipulation of real transactions.
• Do not remove required provenance or AI labels to mislead an audience about the origin of media. Context must make a realistic simulation recognisable as such when it could be mistaken for a real event or recording.

4. Respect materials and privacy#

Do not submit stolen assets, leaked private recordings, confidential work files or material you have no lawful right to use. Licences may differ for a photograph, a composition, a sound recording, a performer, a logo and a building interior. Rights in one element do not automatically clear the others.

LIA is not a service for processing medical records, payment-card details, passwords, government identity records or special-category personal data. Do not use it for covert surveillance, biometric identification or extracting sensitive traits from a person.

5. Harmful and high-stakes uses#

Do not use LIA to facilitate serious crime, terrorism, illegal weapons, exploitation or attacks on critical infrastructure. Do not use creative outputs to make consequential decisions about a person’s employment, credit, insurance, education, healthcare, housing or legal status.

Do not represent a generated scene as a reliable survey, evacuation plan, structural calculation or reconstruction of evidence. LIA is intended for creative exploration, not safety-critical operation.

6. Protect the service#

• No attempts to access someone else’s account, session or private project, including substituting identifiers or reusing stolen login material.
• No malware, credential theft, deliberate overload, unauthorised scraping or circumvention of rate limits, quotas, access controls or provider safeguards.
• No sale or sharing of individual account access and no use of the service to probe systems without the system owner’s authorisation.

7. Reports, review and restrictions#

If an issue is identified, the response may include rejecting a job, limiting a feature, removing material or restricting access. A restriction should be proportionate to the risk and legal obligations. LIA does not claim that every upload is currently reviewed by a human or an automated moderation service.

A report should identify the relevant item or URL, explain the concern and include only the information needed for review. Do not send illegal media as an attachment. Reporting and appeal contact: Information not provided. A working channel and applicable response process must be supplied before public launch.

Where appropriate and lawful, affected users should receive reasons and a route to challenge an error. Reporting a concern does not authorise harassment or false claims. Requests from authorities are subject to applicable law.