Skip to content 
I have just finished the Great Gatsby (again), one of my favourite books of all time, and I am now rereading the wonderful Suitable Boy by Vikram Seth.
Hands Off Our Copyrights: AI, the Copyright Designs and Patents Act 1988, and the Fight for Creative Rights
From healthcare to banking, artificial intelligence (AI) has exploded into practically every industry. However, its growth in the creative industries is one of the most controversial. The emergence of generative AI models that can create text, images, and music at a never-before-seen speed has raised more urgent concerns about copyright infringement and ownership than ever before. A world where machines can “produce” was never intended for under the UK’s copyright laws, particularly the Copyright, Designs and Patents Act 1988 (CDPA). What does that mean for writers, painters, and other creatives, then? And why should we let AI benefit from human creativity without taking responsibility?
We keep hearing this nonsense as to the incredible power of AI but where’s is its subconscious or its soul, it will be found amongst the balance sheet as “follow the money” people. These morons are not interested in anything other than the bottom line. So if its that smart why cannot AI come up with its own art? Its own literature, its own sculpture?
Copyright Protection and the CDPA
The foundation of UK copyright law is the CDPA 1988, which protects sound recordings, motion pictures, broadcasts, and creative works like plays, music, and literature. The author of an original work is automatically granted copyright under this structure, which grants them the sole authority to reproduce, distribute, and modify their invention. The Act makes it apparent that human creativity is at the core of copyright protection by defining a “author” as the person who creates the work.
Nonetheless, there is one significant exception for non-human creators under the CDPA. According to Section 9(3) of the Act, “the author must be understood to be the person by whom the arrangements necessary for the creation of the work are done” in the case of computer-generated works. Although this clause was first intended for algorithmic outputs, like those produced by early computer programmes, its applicability in the era of generative AI is currently up for debate.
The Problem of AI-Generated Works Complex legal issues are brought up by AI-generated content. Who is the owner of the copyright to a painting, an article, or a piece of music produced by AI? Is it the user who enters the prompts, the business that owns the AI system, or the programmer who created the AI model? Some contend that because AI-generated works lack the fundamental human aspect that the CDPA aims to safeguard, they should not be granted copyright at all. Others contend that since AI-generated art is the result of human arrangement, it ought to be protected by current legislation.
The fact that AI cannot “create” in the same manner as humans presents a significant obstacle. It just processes enormous volumes of current data and produces new outputs based on probabilistic modelling; it lacks intent, originality, and artistic vision. This indicates that AI is essentially reliant on human-generated content, frequently being taught on databases that contain copyrighted works without permission.
The Issue of Training Data: A Legal Minefield
One of the most contentious topics in this discussion is the usage of copyrighted content to train AI systems. Large datasets that are scraped from the internet, such as books, articles, music, and artwork, are used to train many AI models, including those created by well-known tech companies. Given that the CDPA gives copyright holders exclusive rights, this presents significant issues.
Text and data mining (TDM), which permits the study of copyrighted works under specific circumstances, mainly for research purposes, is subject to restricted exceptions under UK law. Commercial AI developers, however, frequently take advantage of these restrictions by asserting that their data scraping methods are covered by “fair dealing” or other provisions. Because their creations are being utilised to train AI models without their consent or payment, this violates the fundamental rights of creators.
Copyright Violations and AI: Who Is Responsible?
Who is liable for infringement when an AI-generated work directly mimics or replicates an existing copyrighted work? AI cannot be held accountable for copyright infringements since it lacks legal personality. This places responsibility on the following parties: the AI system’s creators; the businesses using AI for commercial purposes; and the users producing AI outputs.
The person who uses or distributes an AI-generated work may be held accountable under current UK legislation if it is an infringement derivative of a copyrighted work. However, when AI systems function in a “black box” fashion, making it difficult to determine how they arrived at a specific result, infringement might be challenging to prove.
These concerns have already been put to the test in recent court cases where writers and artists have sued AI developers for using their creations in training databases without permission. A precedent that requires AI businesses to seek appropriate licencing agreements—something that ought to have been mandated from the beginning—could be established if such actions are successful.
Legislative Reform’s Need
It is evident from these difficulties that the CDPA 1988 needs immediate revision in order to take into account the reality of artificial intelligence in the creative industries. Possible modifications to the law could include:
- Better Protections for Creators: Putting in place clear rules against using copyrighted content for AI training without the required licence.
- Clarifying the Status of AI-Generated Works: identifying whether AI-generated works qualify for copyright at all and, if so, identifying the legitimate owner.
- Liability Frameworks: Providing precise rules about who is legally liable for infringements of previously created works by AI-generated content.
- Transparency Requirements: These require AI developers to reveal the sources of their training data in order to prevent unauthorised use of copyrighted works.
- Royalty Systems for AI Use: Developing systems that, like current copyright collection societies, provide compensation to creators when their creations are used to train AI.
In summary: Do not Let AI Invade Our Copyrights
AI is a technology that depends solely on human intellectual output to operate; it is not a creator. Without appropriate revision, the rights of authors, artists, and musicians will continue to be undermined because the CDPA 1988’s current legal structure was not created to handle the complexity of AI-generated content. The concepts of originality, justice, and accountability must be upheld by the law.
AI cannot legitimately claim copyright until it is able to produce without depending on the works of actual human creators. And those who create, use, and benefit from AI must be held accountable for infringement if not we will lose.
Stand up People.
Michael.Coyle@lawdit.co.uk
