AI & COPYRIGHT · CASE MAP
AI Copyright Disputes: Evidence Checklist
Separate the dispute into input and training, output, and use or distribution. This shows what evidence a company must retain.
Divide the dispute into three stages
The result may change with the source of the data, terms of service, similarity of expression, human authorship and the way the output is published.
1. Input and training
Record where the work came from, the basis for copying or collection, relevant terms and any access controls.
2. Output
Compare concrete expression, check prompts requesting imitation, and record what expression a person created.
3. Use and distribution
Record whether the output is for internal review, advertising, sale or publication and whether licence terms were met.
Evidence preservation
Keep the service and version, prompts, input files, original source, first output, edits and publication date.
Six questions before release
If the answer is missing, check the source material and contract before publication or sale.
Where did the source come from?
Identify purchase, contract, open licence, employee work or client material.
What do the service terms permit?
Check retention, training, commercial use, responsibility and prohibited uses.
What existing work does it resemble?
Compare concrete expression side by side, not an idea or style alone.
What did a person create?
Keep records of human selection, arrangement, revision and editing.
Where will it be distributed?
State the audience, channel and revenue model.
What can be withdrawn?
Assign owners for takedown, replacement, rights-holder contact and recurrence prevention.
Official sources
Check the linked official text for later amendments and current forms.
Sources checked on 23 August 2026. Copyright analysis changes with the source, contract, output and manner of use.