At the end of 2022, a new artificial
intelligence chatbot was launched and quickly went viral: ChatGPT. The chatbot
created by OpenAI was originally supposed to provide support for online
customer service, but it exceeded all the expectations.
If you ask directly to ChatGPT, it
will answer that its “main function is to generate
human-like text in response to various inputs”. It also explains that it has “been trained on a massive amount of text data”,
which allows it to generate,
not only correct responses, but also contextually appropriate responses.
From
the above, it is clear that ChatGPT generates its responses from textbooks,
newspapers, and any other accessible source of information in order to respond
to queries from the users. However, here
arises one of the IP issues.
As
this chatbot can generate texts, the concern is who is the author of the
content, but also, if it has the protection the human creators receive from
their creation. First, it is important to point out that a work is considered
as copyright and it gives protection to its author as soon as it is created in
a fixed tangible form. This means that the content shall be available in a
physical or digital form to be protected by copyright law.
However,
in the case of ChatGPT, it is unclear whether the created content can be
protected by copyright law. While the created content could be protected by
copyright, it will not be owned by the AI itself. In that regard, the Mexican, European
and US laws state that AI cannot own copyright, as it does not have the legal
personality, which is a pre-requisite to own such asset, and thus, it cannot be
recognized as an author.
In
light of the above, ChatGPT cannot own what it creates, but the responses could
be protectable by copyright. Although it has not yet been settled by the law,
it seems that creators who use AI to support their work may be able to claim
ownership of their work, only if they are able to evidence that the work is the
result of their creativity. However, if a user asks the chatbot to create a
song without more information, it is unlikely that the user would be able to
claim ownership over the song.
On
the other hand, it is important to point out that the content created by
ChatGPT derives from several data, that is, information that was generated by
others. Thus, by using such content, you could also infringe authors´ rights.
In
that regard, some artists were concerned by the use of ChatGPT, as it has been
trained on their works.
Recently,
some companies, as GETTY IMAGES sued OpenAI for training on millions of its
pictures without consent, and several artists have expressed the same concern.
These cases will be litigated in the United Kingdom, but it is expected that
artist from all around the world will express the same concern.
In
the United States, several authors filed a lawsuit against OpenAI and its
chatbot for copyright infringement and violation of other laws. They claimed
that the chatbot was trained on their books, without consent, and can therefore
generate summaries of the author´s books, which is a clear infringement of
their rights.
These
lawsuits are currently pending, but the way in which they are ruled would
certainly set up some precedents for AI.
Another issue with ChatGPT is that
several users have been using AI for creating intellectual property. Indeed, it
has been used to draft patents, cease and desist letters, contracts and any
other legal document. Besides the fact that it is not recommended to use AI for
such purpose, as the information is publicly available (which could ruin
novelty if we talk about patents), the AI cannot replace the legal expertise of
an attorney specialized in IP, who has valuable knowledge acquired from his
experience.
In addition, ChatGPT may not take
into consideration the country where you live, and the response may be
applicable only for some jurisdictions, in which the applicable law differs
from where you are.
Although ChatGPT seems to be an
interesting tool in many fields, including intellectual property, users shall
not only rely on AI to obtain information and assistance regarding their IP
rights, as the provided response may not take into consideration the
jurisdiction, but worse, it could make confidential information become public.
Vanessa Rondeau Galindo