The Law of All Things

The use of AI Platforms within business has been a hot topic for some time, and the launch of ChatGPT by OpenAI at the end of 2022 has exploded the discussion about how it can be utilised, and the possibilities are endless.

But behind the excitement and fervour, it’s vital for Companies, Business Owners and Entrepreneurs alike to understand the legal implications around their use, and more specifically, who owns the content produced by Artificial Intelligence Platforms?

In the EU, under the Database Directive, the works must be the author’s intellectual invention to benefit from copyright protection at the EU level. In the UK, any literary, musical, or artistic work must be ‘original’. In the U.S it must be a result of original and creative authorship by a human author.

In the past, the court interpreted that to mean that the author must have created the work using their talent, judgement, and effort. The human element is ever-present, and the question of whether AI is capable of meeting the threshold for originality is central to any copyright-AI discussion.

In the 2021 case of Stephen L Thaler v. Comptroller-General of Patents, Designs, and Trade Marks, the UK Court of Appeal determined that a machine is not an inventor for the UK Patents Act 1977. Nonetheless, the underlying copyright and patent positions on ownership of computer-generated works remain unclear. However, even if copyright applies to an AI-generated work, the ownership issue remains.

The ownership of copyrighted works produced by AI systems is a crucial issue in copyright. The method by which UK copyright legislation, such as the Copyright, Designs, and Patents Act 1988, deals with ‘computer-generated’ works is complicated. A computer-generated work is owned by the person who makes the arrangements required to create it under UK law. This has been referred to as the “human author” who created the piece using the programme.

However, there is no guarantee that this would automatically incorporate new technologies and works that may be produced in new, unanticipated ways. The aforementioned clause could be interpreted to only apply in circumstances where the deemed author, the necessary arrangements, and the creation of the work are all close by, leaving out instances in which the final product is created very far down the line independently from the person or even without the person’s presence. To put it another way, designing or making a piece of equipment has nothing to do with how a person uses it or what they produce from it, let alone when the equipment is acting independently and creatively. This should not come as a surprise, as those laws and regulations were made back when computers still used punched cards. Therefore, AI was not at the forefront of lawmakers’ minds at the time.

In an October 2021 consultation document, the UK government proposed making changes to UK IP law in three important areas:

“Without a human author, computer-generated works (CGWs) are protected by copyright; text and data mining (TDM) is licenced or exempt from copyright; and AI-developed ideas are protected by patents.”

Regarding CGWs and innovations created by AI, the UK concluded that no adjustments were required in its answer to the consultation, which was released in June 2022. This was done for two reasons: artificial intelligence is still in its early stages, and global harmonisation is critical.

Does OpenAI own the copyright to ChatGPT outputs?

AI solutions, such as ChatGPT, are unlikely to produce original works as their outputs are produced using a combination of pre-existing data and the model’s learned weights and biases. Therefore, for the time being, it is unlikely for these to be capable of being copyrighted by the AI itself.

OpenAI dealt with this matter in a very clever way. As stated in their Terms of Use, clause 3(a):

“Your Content. You may provide input to the Services (“Input”), and receive output generated and returned by the Services based on the Input (“Output”). Input and Output are collectively “Content.” As between the parties and to the extent permitted by applicable law, you own all Input, and subject to your compliance with these Terms, OpenAI hereby assigns to you all its right, title and interest in and to Output. OpenAI may use Content as necessary to provide and maintain the Services, comply with applicable law, and enforce our policies. You are responsible for Content, including for ensuring that it does not violate any applicable law or these Terms.”

Essentially, this means that, subject to the relevant laws, OpenAI assigns to the user all its rights, titles, and interests in and to the created content. As a result, the user owns the output generated by GPT models, subject to the terms and conditions of the licence agreement and compliance with all applicable laws.

On the other hand, AI is just a tool, and the person controlling the AI should be allowed to claim ownership of the output.  However, when it comes to ChatGPT, the person’s control over the output is limited. Therefore, an argument could be made that the output is more in the hands of the ChatGPT software’s developer(s) than it is of the user who begins an input.

Returning back to OpenAI Terms of Use, clause 3(a):

Similarity of Content. Due to the nature of machine learning, Output may not be unique across users and the Services may generate the same or similar output for OpenAI or a third party. For example, you may provide input to a model such as “What color is the sky?” and receive output such as “The sky is blue.” Other users may also ask similar questions and receive the same response. Responses that are requested by and generated for other users are not considered your Content.”

ChatGPT can generate the exact same output for various users. In a copyright infringement claim, evidence of copying is necessary. However, the idea of independent creation may prevent two persons whose queries produced the same work from being able to defend their rights. Here, neither user would have replicated the other’s work and as such no infringement action is likely to be successful.

Contracts are more important than ever

Because the laws around the world aren’t clear on the matter, the parties to agreements for the creation and use of AI systems that may lead to new, intellectual property-protected works should include rules about ownership, assignment, and licensing.

Similarly, the application of AI systems may result in patentable innovations. As such, contractual parties should once again specifically make provisions in their AI contracts for the ownership, assignment, and licencing of inventions and intellectual property rights created by AI.

The UK High Court denied a patent application for a machine-created innovation in September 2020 because an “inventor” had to be a natural person. The issue was also brought up before the US Patent and Trademark Office, which similarly denied the award of patent applications in the US, and the European Patent Office, which rejected two European patent applications claiming an AI system as the inventor. On a patent, only real people can be listed as inventors. Every court has come to this conclusion.


How can Fusion Law help?

If your company uses AI to produce, use, or generate content, it is important to explore contractual protections or restrictions that may be necessary to safeguard your position, especially until the legislators provide further clarity.

Back to Blog