Copyright and databases – who owns what, if AI is the creator?

AI is here to stay. Recent developments in AI have shown how much it has advanced in the last decade (we promise that Chat GPT did not write this article!) At the moment, however, we are currently in no man’s land with regard to ownership of copyright and database rights when such works are created by AI. The Government’s Science, Innovation and Technology Committee have recently released an interim report titled ‘The governance of artificial intelligence’. Whilst the report is useful in highlighting the worries of copyright owners, it does not provide any substantive answers as to how the Government are going to deal with the problem of AI infringing copyright.

Under UK law, copyright is legislated by the Copyright, Designs and Patents Act 1988 (CDPA 1988). For copyright to subsist in a work, it must fall into one of the categories found within the CDPA 1998 and qualify for protection. The CDPA 1988 protects original literary, dramatic, musical and artistic works (LDMA) as well as typographical arrangements, broadcasting, films and sound recordings, called the ‘entrepreneurial works’. The reason for this distinction is to protect the creators of the work via ‘classic’ copyright protection, and investors in such works (e.g. produces) via ‘entrepreneurial’ copyright protection.

Subsistence

To qualify for copyright protection, LDMA works must first be original. Under classical UK law, a work being ‘original’ meant that the work had to be created through the author’s own skill, judgment and individual effort whilst not being copied from another work (Ascot Jockey Club Ltd -v- Simons [1968] 64 WWR 441). However, recent UK cases have used the EU’s approach to originality which says the work must be of the “author’s own intellectual creation” (Infopaq International A/S v Danske Dagblades Forening (Case C-5/08). Case law (Cofemel v G-Star Raw (C-683/17)) has clarified what is meant by the “author’s own intellectual creation” with two elements being established:

1. The subject matter concerned has to be original in the sense that it was the author’s own intellectual creation. In Cofemel, the court said that it was both necessary and sufficient that the subject matter reflected the personality of the author as an expression of their free and creative choices. This is known as the originality requirement.
2. The object in question which is deemed to have been an expression of the author’s own intellectual creation has to have been able to be classified as ‘work’ within the Copyright Directive. This is not a problem in the UK as we function on a closed list, for example the work has to fit into one of the established classical or entrepreneurial rights. This element is known as the identifiability requirement.

There is a difficult question to be answered whether a work created by AI has it’s own intellectual creation which, therefore, impacts ownership of the work. The answer, at the time of writing, remains unknown as the Court has not handed down a decision.

In the writer’s view, it is unlikely that copyright would subsist in a work created by AI. This is because the AI itself cannot independently create the work; it needs to be fed information which it uses to create the work. Therefore, how can it come up with independent thought and thus works? The answer is that it cannot.

Ownership

Section 11 of the CDPA 1998 says that the first owner will be the author. Author is defined under s.9(1) of the CDPA 1988 as the person who creates the work. As with other intellectual property rights, where work is created during the author’s employment, the employer will own the copyright in the work. However, if a work created by AI meets the originality requirement, which, as discussed above, in the writer’s view is unlikely, who would the author be of a work created by AI?

Unlike many jurisdictions such as the EU, the US, Australia and China, the CDPA 1998 expressly provides protection to works which are completely computer-generated. s.9(3) of the CDPA 1998 says that for a work which ‘is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken’. s.178 of the CDPA 1998 defines ‘computer-generated’ to mean works generated by a computer where there is no human author. If the legislation means that the behind the AI will be the author, then how much intellectual input did that person have? Was there any ‘intellectual creation’ from this person? It is likely that s.9(3) of the CDPA 1998 means the person who inputted the instructions to the AI will be the author, however, at the time of writing this has not been confirmed by a Court.

On the other hand, entrepreneurial rights, as previously mentioned, do not need to be original. Therefore, in the writer’s view, these types of work are more likely to receive copyright protection. For example, if AI created a song which it then recorded, the individual who inputted the necessary information for that AI to create the song would likely be the author. Whilst this hasn’t been confirmed, as there does not have to be an intellectual creation on behalf of the AI, the threshold is lower.

Database rights 

Whilst copyright can subsist in databases, database rights are a separate right under the UK intellectual property (IP) law. Databases are protected by the Copyright and Rights in Databases Regulations which was enacted in 1998 Regulations.

Databases are defined in the Regulations as ‘a collection of independent works, data or other materials which (a) are arranged in a systematic or methodical war, and (b) are individually accessible by electronic or other means’. Usually databases take the form of things such as contract management systems, intranets, knowledge management systems, mailing lists and/or inventory systems.

Subsistence

To qualify for protection as a database, there must be substantial investment in obtaining, verifying or presenting the contents of the database. This is an intentionally low threshold to ensure that smaller databases are protected rather than just large ones (British Horseracing Board Ltd and others -v- William Hill Organisation Ltd).

It is trite law that the owner of the original data used to create the database will not benefit from database protection. Therefore, a database created by AI, in theory, could qualify for database rights provided that there was ‘substantial investment’ in creating the database. There has been no case which defines what substantial investment would be with regard to AI but this could mean substantial investment by either the individual feeding the information to the AI or the creator themselves.

Ownership

The owner of a database is the maker. The ‘maker’ is the person who takes the initiative in obtaining, verifying or presenting the contents within the database and assumes the risk of investing in the database. As is common with IP, where an employee creates the database during the course of their employment, the employer will own the database.

However, the Regulations are silent on who would own a computer-generated database. If a database created by AI does constitute ‘substantial investment’, the owner would likely be the person who fed the AI the data. However, it is to be determined how a Court would pass judgment on this issue. It may be the person who requested that the database be created. However, the question is therefore whether that person would have to input substantial investment.

Until a case is heard in the UK Courts, these queries will remain.

For more information, please contact Chris Fotheringham or Jenny Guild.