Do AI Technology Providers Infringe Copyright? 


In the rapidly evolving world of artificial intelligence, new questions are arising regarding copyright and intellectual property. The advent of advanced language models like ChatGPT has sparked concerns amongst several authors, including renowned writers John Grisham and George R.R. Martin. They have initiated lawsuits against ChatGPT-owner OpenAI, claiming that their works were used without their permission to make ChatGPT smarter. OpenAI said in response that it respects the rights of authors. 

Understanding Copyright Protection in the UK 

In the UK, copyright protection is primarily governed by the Copyright, Designs & Patents Act 1988 (“CDPA”). The CDPA protects original literary, dramatic, musical, and artistic works. It grants exclusive rights to the authors of such works, allowing them to control the reproduction, adaptation, and distribution of their creations. 

The UK is also a signatory to several international copyright conventions, including the Berne Convention for the Protection of Literary and Artistic Works and the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS). These conventions establish minimum standards of protection for copyrighted works globally. 

While international copyright conventions provide a framework for copyright protection, they do not specifically address AI-generated content. The interpretation and application of these conventions to AI language models represent a legal landscape that is yet to be fully defined. 

Examining AI-Generated Content 

ChatGPT and other large language models “learn” by analysing a massive amount of data often sourced online, known as training data. They do not produce original works in the traditional human sense, but rather present outputs that mimic existing content. This poses unique challenges when evaluating copyright claims against AI providers. 

One of the central concerns in copyright claims against AI language models is determining the ownership and attribution of the generated text. The CDPA stipulates that copyright belongs to the author or creator of a work. However, in the case of AI-generated content, the question arises as to whether the author is the human operator or the AI system itself. 

Under the CDPA, for a work to be protected, it must be original and demonstrate the skill and labour of the author. As AI models like ChatGPT do not possess the inherent creativity and intent of human authors, establishing authorship becomes a complex matter. This poses a challenge for claimants seeking to establish their copyright ownership over AI-generated content. 

Substantial Similarity and Infringement 

Determining infringement also requires establishing substantial similarity between the AI-generated content and the copyrighted work. In the case of AI language models, the training data often consists of diverse sources, making it difficult to establish a direct connection to any single author’s work. 

The CDPA’s test for substantial similarity involves assessing the quality and quantity of the copied elements, as well as the importance of those elements in the original work. It remains uncertain whether courts would consider AI-generated content to be substantially similar to any specific copyrighted work, given its algorithmically synthesized nature. 

The CDPA also notably includes a provision for fair dealing, which allows limited use of copyrighted material without explicit permission. Fair dealing enables actions such as criticism, review, news reporting, research, and private study. It remains to be seen how these fair dealing provisions may apply to AI-generated content, and whether they can be invoked as a defence in copyright claims against AI providers. 


The outcome of copyright claims against the likes of ChatGPT will shape the legal landscape surrounding AI-generated content. A favourable ruling for claimants may lead to stricter regulations on AI models, or the implementation of content filters to mitigate potential copyright infringement. 

However, the intricate nature of AI-generated content, challenges in establishing ownership and originality, and the issue of substantial similarity pose significant hurdles for claimants. If such copyright claims prove unsuccessful, it may reinforce the idea that AI-generated content lacks the necessary human creativity and intent for copyright protection.  

Such an outcome would likely encourage the continued development and deployment of AI language models for various purposes, facilitating further innovation. As AI progresses, the courts will likely seek to strike a balance between copyright protection and the transformative potential of this remarkable technology. 

*This article is not legal advice but provides a general overview. The specific details of your case will determine the best course of action.