In the third of a series of articles looking at how Generative AI is impacting trade marks and designs, we look at the legal implications of ownership when it comes to AI generated content. (For further information, and to view the other articles in the series, please visit our Generative AI hub).
As artificial intelligence (AI) continues to blur the lines between human and machine creativity and advance in its ability to create content, it raises significant questions around ownership and intellectual property (IP) rights.
Current position in the UK
The default position is that copyright in a literary, dramatic, musical or artistic (LDMA) work, under the Copyright Designs & Patents Act 1988 (CDPA), is owned by the ‘human author’. Specifically, in the case of a computer-generated artwork[1], the author will be “the person by whom the arrangements necessary for the creation of the work are undertaken”. UK law allows for copyright in these ‘computer-generated works’, with a reduced term of 50 years from when it was created (as opposed to the 70 years from the death of the author in the case of content created by a human).
It is not immediately obvious who the ‘human author’ is for AI-generated content. There is scope for it to be any of:
- the user of the generative AI tool who enters the inputs into the software;
- the writer of the generative AI software;
- the person that taught the generative AI software; or
- the owner of the copyright in any material used to teach the generative AI software and to create the new content.
Another obstacle is that under the CDPA, any LDMA works eligible for copyright protection must be original. Using an example of ChatGPT, which enjoys 82.5% of all AI tool website traffic[2], information and sources used to teach ChatGPT is from publicly-available information on the internet and user-inputted information[3] – the underlying work may therefore still be protected from the original copyright.
The standard for ‘originality’ has historically been rather low, where ‘original’ simply meant ‘not copied’. However, recent case law[4] has confirmed that the ‘author’s own intellectual creation’ is required; a higher standard because it introduces a creative element. The author will be the person who contributes the creative skill and labour. The writer of the software clearly qualifies in this context but it could also be the person who originated the expression of the ideas in the work.
Effect of terms and conditions of generative AI applications
Most AI tools have terms and conditions that affect ownership of the generated content. While some allow users to retain ownership of outputs, others have provisions that may limit users' rights.
For instance, tools like MidJourney allow users to own outputs if they are paying subscribers[5]. Free-tier users may not have exclusive ownership, as content can be reused or shared by others. On the contrary, the usage policies for OpenAI’s ChatGPT state that the user owns the rights to the content they generate[6]. However, it is important to consider that these Terms of Use also grant OpenAI the right to use user-generated content for improving the model, conducting research, or ensuring compliance with their policies.
Potential future position in the UK
As the UK has yet to have any AI-specific regulation or legislation covering generative AI tools, it is instead regulated through existing legal frameworks.
The UK government and UK Intellectual Property Office continue to hold open consultations to “ensure the UK gets the national and international governance of AI technologies right”. In 2022, the feedback from one of the first consultations indicated that the tech sector believes the copyright to AI-generated content should belong to users or developers, whereas the creative industries want this content to be excluded from copyright protection. This was followed up in 2023 by a White Paper on AI Regulation[7], setting out the Government’s pro-innovation and principles-based approach to AI regulation. The UK Government has not yet acted on the findings and instead recommended further consultation so it is unclear what action will be enforced at this stage.
Conclusion
Finally, as an important viewpoint to consider, I asked ChatGPT for its view on who should ultimately own content produced by generative AI tools, and it said, “Ownership should ultimately reflect a balance between user contributions, the technology's role, and societal interests. This ensures fairness while fostering innovation and responsible use of AI. Governments and regulatory bodies need to establish clearer laws on AI-generated content to address ambiguities in copyright and ownership”. ChatGPT helpfully identifies the challenge but doesn't come up with a solution. Human policy and law makers will have to manage that on their own, it seems.
[1] Section 9(3) of the Copyright Designs & Patents Act 1988
[3] https://help.openai.com/en/articles/7842364-how-chatgpt-and-our-foundation-models-are-developed
[4] THJ Systems Ltd v Sheridan [2023]EWCA Civ 1354
[5] https://docs.midjourney.com/docs/terms-of-service
[6] https://openai.com/policies/row-terms-of-use/
[7] https://www.gov.uk/government/publications/ai-regulation-a-pro-innovation-approach/white-paper
