ChatGPT Causes Copyright Concerns

ChatGPT Causes Copyright Concerns

ChatGPT has taken the world by storm. It is the fastest-growing consumer application ever launched, achieving 100 million active users two months after its release. Users are drawn to the tool’s sophisticated capabilities but worried about its potential to fundamentally change numerous industries. A much less discussed implication is the privacy risks ChatGPT poses to each and every one of us. Google only yesterday launched Bard, its own conversational AI, and others will undoubtedly follow. Technological firms engaged in AI development have clearly entered a race against each other.

How does ChatGPT work?

ChatGPT is a language model AI computer program which generates output in the form of dialogue from text-based input received from the user. The output produced by the ChatGPT chatbot can be in a variety of formats, such as essays, research papers, and even source code. As a result, a user can ask the chatbot a question and get a response that sounds human-like. Large language models that are the foundation of ChatGPT require enormous amounts of data to operate and develop. The more data the model is trained on, the more proficient it becomes at seeing patterns, foreseeing what will happen next, and producing credible language.

An example of prompting ChatGPT for a response.
Source: SpringBoard

The startup behind ChatGPT, OpenAI, supplied the programme 300 billion words that were systematically scraped from the internet, including postings, books, articles, webpages, and other sources of unapproved personal information. There is a strong possibility ChatGPT has read something you said in a blog post, a product review, or a comment you made online.

Scraped data that ChatGPT was trained on can be proprietary or copyrighted. For instance, when prompted, the tool produced the first few passages from Joseph Heller’s book

Catch-22 – a copyrighted text.

ChatGPT does not necessarily consider copyright protection when generating outputs.
Source:TheConversation

It is unclear who can copyright or claim ownership of AI-generated works. The requester, who simply used a tool to generate text, or OpenAI? Who?

Margaret Esquenet, partner with Finnegan, Henderson, Farabow, Garrett & Dunner, LLP, states that for a work to enjoy copyright protection, the work must be the result of original and creative authorship by a human author. A work is not entitled to copyright protection if there is absence of human creative input. Currently, the U.S. Copyright Office will not register a work that was created by an autonomous artificial intelligence tool.

Additionally, this issue is currently the subject of litigation, for example, in a case involving a photograph shot by a monkey and lawsuits examining the idea of AI inventorship in the context of patents. The idea of non-humans claiming authorship or invention, and in both cases, consequently ownership of the intellectual property (IP), has thus far been met with hostility by the courts.

Assuming the prior cases forbidding authorship for non-humans are upheld, no one would be able to possess the product thereby donating such works to the public. As the vast majority of the source materials are at least readily accessible to the public, such conclusions could also be supported. As a result, using ChatGPT to produce academic articles, research papers, or the like raises major ethical and academic integrity issues.

Who would be the creator and proprietor of the copyrighted work produced by such an artificial intelligence, then, in terms of intellectual property? South African copyright law distinguishes between the author and owner of a copyright work, as the author of the work may not necessarily be the owner thereof.

ChatGPT’s terms of use clearly assigns all its rights, title and interest in and to the output or content created, to the user. However, does ChatGPT even own the content which it so readily assigns? According to Act 98 in the Copyright Act of 1978, the author is normally the first proprietor of any copyright work, but there may be exceptions to this rule that are outlined in the Act.

Depending on the type of work produced, ChatGPT output will have a variety of authors. A literary work’s author, for instance, is the person who first makes or creates the piece, but a computer program’s author is the person who has control over the program’s creation. In light of this, it is important to consider who would have written an essay if a computer programme had generated it.

The court made a distinct distinction between computer-generated work and computer-aided work in Payen Components South Africa Ltd v. Bovin Gaskets CC. Work produced by a computer programme that was invented and designed by a human author is referred to as computer-generated work. However, in the case of work generated by AI, the work is clearly not generated by a human author and control and direction is expended by AI.

A compelling argument may be that AI is simply a tool and that the human who is directing the AI should be able to claim ownership of the output. For example, a graphic artist can claim artwork made through the use of drawing software. However, in the case of ChatGPT, the operator’s control of the output is limited, and perhaps a stronger argument could be made that the output is controlled more by the creator of the ChatGPT software than the operator who initiates an input

What about when ChatGPT generates the exact same passages for someone else?

Even if the individual whose queries produced the AI work is the legitimate copyright owner, the idea of separate creation may prevent two persons whose queries produced the same work from being able to enforce rights against one another. In particular, independent creation is a complete defence to a copyright infringement accusation; proof of copying is necessary. Since neither side in this example replicated the other’s work, no infringement action is likely to be successful.

However, if the content created is damaging, who gets the blame? An exact copy of protected work could create potential liability, which raises another question: who is liable, the creator of the AI — such as ChatGPT — or the user who posed the query? If generated text is used in an article or paper, even partially, should ChatGPT be treated and cited as a source in itself? As in “ChatGPT-generated response, accessed 12/15/2022”? Correctly citing AI-generated content is okay. From a Blue Book citation standpoint — citing in court briefs — as well as for literary purposes, a citation to the source as ChatGPT would be appropriate.

Getty Images filed a lawsuit in the High Court of London in January 2023 against the AI company Stability AI. Stable AI allegedly copied millions of photos from Getty Images’ database to train its image generating model. Commentators have suggested that Stable AI’s photographs may have been used to train its model by pointing out that they frequently bear the watermark from Getty Images.

An original Getty photograph, left, an AI image generated by Stable Diffusion complete with Getty logo on, right.
Source: 
PetaPixel 


Although it is unclear how much ChatGPT’s model will change the original works to prevent infringement, the Stability AI example raises serious concerns about the usage of AI that has been trained on publicly available works and may even pose an existential danger to some models. Some customers, especially commercial users, may be hesitant to employ ChatGPT or equivalent AI in their organisation because of the danger of infringement.

As noted above, the infringement risk extends to the user and not just the platform.

Application in IP law

ChatGPT’s development is undoubtedly another revolutionary step in the field of artificial intelligence, but it also raises more questions about how IP law, particularly copyright law, should be applied. The identity of the author and owner of the output, despite an assignment provision contained in its terms of use, is still open to speculation and accordingly creates copyright works in which proprietorship cannot necessarily be identified.

This is indicative of the necessity for copyright law to adapt to the latest technology. It is vital for an individual or business to gain exclusive rights to their creative work. Learn how you can register your copyright with us here at Intellect.