KazakhstanLaw on Artificial Intelligence, copyright and training data
In force since .
An AI training data rule binding public and private bodies.
- Audit expectation
- none
What it requires
- A work an AI system produces is protected by copyright only where a human made a creative contribution to it.
- Do not train an AI model on a published work if the author or rightsholder has posted a prohibition against it in machine-readable form.
- Training a model on a work is not covered by the educational or scientific free-use exception in Kazakhstan's copyright law, so it needs its own basis under this Article rather than that exception.
These lines are LexLint's reading of the law's text.
In the frameworks
- NIST AI RMF
- MAP 4.1Approaches for mapping AI technology and legal risks of its components – including the use of third-party data or software – are in place, followed, and documented, as are risks of infringement of a third party’s intellectual property or other rights.
Our reading of where this law's duties sit in each framework. It never says that running a control meets the law. How we read a framework against the law
What it makes you log
Logging duty
Article 23 sets copyright and training-use rules (human creative contribution, the opt-out condition expressed in machine-readable form) and states no duty to log, record or document a training run, a prompt, or a check for an author's or rightsholder's machine-readable prohibition.
- Kind
- None
- As of
What this law does
Article 23(1)-(2) protects a work created using an AI system by copyright only where there is a human creative contribution to its creation, and recognises a text prompt sent to an AI system as a copyright object where it is itself the result of a person's intellectual creative activity.
Article 23(3) states that using a work to train an AI model does not fall within copyright law's free-use exceptions for educational or scientific purposes, and Article 23(4) states that training does not itself amount to exercising the author's reproduction, distribution, adaptation, public display, performance, broadcast or cable-communication rights.
Article 23(5) then permits using a work to train an AI model only where the author or rightsholder has not expressed a prohibition in machine-readable form, an opt-out condition on top of the two preceding paragraphs.
When LexLint raises it
When your app profile says your app trains models or generates content with AI.
Back to the example · Lint your app
Page updated .