Have you ever wondered who owns the content that you prompted and which is then provided by ChatGPT or other AI-based writing tools? Who owns the copyright? How do big players like Google or Microsoft handle this situation?
But first things first. Almost every country has published its own copyright law or regulation regarding AI. Thus, it makes it even more difficult to figure out who actually owns a certain piece of art or article or invention.
The Digital Watch observatory provides a very interesting overview on AI and copyrights.
It states that there exist basically two camps (US, EU (including Austria), etc.): the first one says that nobody owns the results which have been produced by AI because ‘copyright applies exclusively to original works requiring that originality represents the author’s own intellectual creation. This typically means that an original work must reflect the author’s personal input, highlighting the need for a human author for copyright eligibility.’
In other words, only works produced by human beings can have intellectual property rights and any output produced by AI is free for everyone to use. So far, this approach has been adopted by most countries.
Nevertheless, the second camp (UK, India, Ireland, New Zealand, etc.) attributes authorship to the human beings involved in the production process such as programmers or developers.
The CDPA of UK’s copyright law says: ‘In the case of a literary, dramatic, musical or artistic 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.’
Unfortunately, none of the above helps you to find out who the legitimate owner or author is. The vast amount of information that has been attained through web scraping and then used to train these so-called LLMs (Large Language Models) makes it almost impossible to figure out the original sources. But in case some output can be traced back to a copyright protected piece of work, it depends on the country whether you can file a suit against the copyright infringer.
That’s why large companies like Microsoft have launched their own initiatives such as the Copilot Copyright Commitment where Microsoft says that ‘The Copilot Copyright Commitment extends Microsoft’s existing IP indemnification coverage to copyright claims relating to the use of our AI-powered Copilots, including the output they generate, specifically for paid versions of Microsoft commercial Copilot services and Bing Chat Enterprise.’ and promises to ‘pay for any adverse judgments if they are sued for copyright infringement for the use of the Azure OpenAI Service outputs’.
Google catched up on that in October 2023 and announced in regards to their Duet AI – their always-on AI collaborator – used in Google Cloud and Workspace: ‘if you are challenged on copyright grounds, we will assume responsibility for the potential legal risks involved.’
Good to know, Google guarantees help not only for the output generated by its products, but also for its use of training data.
In Austria, §1 of the copyright law says that you can only protect creations that are the result of human activity and a software program cannot be viewed as the author. Regarding the documents used to train LLMs which have mostly been created by human beings, §42h on text and data mining of the Austrian copyright law allows the use of copyrighted works to train an AI as long as the copyright owners do not expressly object to this using machine-readable means. Additionally, if you want to sue someone for copyright infringement, you have to prove that the AI used had access to the document protected by copyright.
There are still so many open questions and I’m quite sure that this topic will stick with us for many years to come. Hopefully, governments all over the world will see these initiatives as wake-up calls and gather the involved stakeholders to rethink the current approaches to IP rights regarding AI 🤞