Legal Position

Fully cleared. Fully indemnified.

At Lingotion, legal clarity is part of the product – not an afterthought. Every Character you license is built on properly consented, properly documented, and properly compensated performance data, so you’re not exposed to the copyright, likeness, or data-provenance risk that comes with unlicensed AI training data.

The five legal issues with generative AI

Before explaining what we do, here’s the framework we use to evaluate legal risk in any generative AI system – the same lens we’d encourage you to apply to any vendor, including us.

Copyright
Do you have the right to train on the data?
Could the output infringe a third party’s copyright?
Does the output itself qualify for copyright protection?
Acquiring the data
Under what legal terms was the data acquired?
Personal data
Can a real individual be identified in the data?

How Lingotion answers each one

  1. Right to train: every Recording is covered by a direct, written actor agreement that explicitly licenses Lingotion to train AI models on it – never scraped or third-party data.
  2. Third-party infringement in output: because models are trained only on licensed, consented Recordings, there’s no public scraped dataset in the mix that could surface someone else’s copyrighted material in the output.
  3. Copyright on the output: most generative AI output isn’t copyrightable – courts have generally held that a prompt alone doesn’t make a human the author. Lingotion trains its models to stay derivative of the original licensed performance rather than transformative, so the actor’s copyright in that performance is designed to carry through into the AI output. We keep training logs documenting that balance for compliance purposes.
  4. Terms of acquisition: each actor agreement sets out in writing what was licensed, for how long, and under what compensation – see the Actor Payment terms in our Terms of Service.
  5. Personal data and identifiability: actor consent, chosen Ethics Levels, and our GDPR-based Privacy Policy govern how identifiable data in Recordings is used.

Full provenance, no scraped data

All Lingotion voice and performance models are trained exclusively on data we have the explicit right to use for AI training.

  • No data scraped from the internet
  • No use of content under terms that don’t permit AI training
  • No public models trained on undocumented or unlicensed data
  • Full compliance documentation available to paying customers on request

You're covered: Lingotion's indemnification

When you license a Character, Lingotion contractually indemnifies you and your Legal Group against claims arising from:

  • Any breach of the Agreement by Lingotion
  • Lingotion’s violation of a third party’s IP or privacy rights
  • Lingotion’s violation of any law, rule, or regulation
  • Actor payment obligations
  • Actor’s legal rights

In practice, the legal risk of using a Character sits with Lingotion, not your studio – as long as you use it within the terms of your license.

Actor consent
→
Lingotion ownership
→
Indemnified license
→
Your studio

What your license actually gives you

Lingotion owns the underlying IP in its Recordings, Characters, and technology – that ownership is what lets us guarantee clean provenance and stand behind the indemnification above. What you receive is a clear, non-exclusive license to use the Character within your agreed Market Segment (Video Games, Education, Product Marketing, Entertainment, or Internal Use) and Ethics Level, for the Project you’ve licensed it for.

  • License terms (Market Segment, Ethics Level, Project scope) are fixed at purchase and shown in your portal
  • Using a Character outside its licensed scope requires written approval from Lingotion and, where relevant, the Actor

Your license is protected by the indemnification above for the full term of use.

Actor consent and compensation, by design

Every Character is built on top of explicit actor consent, with usage rights defined upfront and compensation tied to actual use. Actors set their own allowed Market Segments and Ethics Level, and nothing outside those bounds happens without their sign-off. That consent chain is part of what makes the legal position above possible – you’re licensing performances that were properly cleared from the start.

“If you cannot prove the data chain, you carry the risk.”
Andreas Rodman, Founder of Lingotion

Read the full terms

This page is a plain-language summary. The complete, binding terms – including Market Segment definitions, Ethics Levels, indemnification, and liability terms – are set out in our Terms of Service.

Questions about how this applies to your project? Reach out to sales@lingotion.com.