1) Why should OpenAI fear a Scarlett Johansson lawsuit?
(Por que a OpenAI deveria temer um processo de Scarlett Johansson?)
OpenAI’s concern about a possible lawsuit filed by actress Scarlett Johansson pertains to a formal request from her lawyers for the company to remove the “Sky” voice option from the new ChatGPT-4o app model because that voice was deliberately similar to Scarlett’s. The discussion revolves around topics such as the protection of personality rights, including the right to a voice, and intellectual property issues pertaining to the development and use of Artificial Intelligence technologies. It is worth bearing in mind that this agenda was the subject of discussion during the Hollywood writers’ and actors’ strike in 2023 and, in the US scenario, there is regulation of the matter through general principles set out in the “Blueprint for an AI Bill of Rights” and in the “Artificial Intelligence Risk Management Framework,” prepared by NIST (National Institute of Standards and Technology). In addition, President Joe Biden recently issued a new Executive Order on Secure and Trustworthy Artificial Intelligence, establishing guidelines for companies developing advanced and higher-risk AI technologies, such as generative AI, to which ChatGPT belongs. It is also worth pointing out that there are rules on the subject also in force in the United Kingdom (IA Act) and, in Brazil, we have EBIA – Brazilian Artificial Intelligence Strategy. These are some of the reasons why OpenAI should fear a lawsuit from actress Scarlet Johansson.
__________________________________________________________
2) Does Scarlett Johansson have a fair claim? Why?
(Scarlett Johansson tem uma reivindicação justa? Por que?)
Although we do not have access to the lawsuit in question, it is reasonable to state that actress Scarlett Johansson does have a fair claim. From the standpoint of facts, OpenAI itself stated in a clarification note published on its official website, entitled “how the voices of ChatGPT were chosen” that the company’s CEO, Sam Altman, contacted the actress to discuss her participation as a voice of one of the voices of the new ChatGPT voice mode, a proposal that was rejected by the actress. In addition, Sam made posts promoting the new model on his X social network profile referring to the film “HER,” in which the actress lent her voice to a voice assistant, the central character of the plot, as well as the chatbox “Sky,” voice resource of the GPT-4o model. In the legal perspective, the voice is considered as a sound extension of the personality, being an important individualizing and identifying element of the person in the social body, and, also for this reason, it receives specific protection, including copyright related to its commercial use, both in the Brazilian and in the US contexts. In addition, it also supports the claim that OpenAI’s models are not transparent and, seemingly, have been trained with copyrighted works, even if published on the internet, a situation similar to that discussed by The New York Times in a lawsuit filed against OpenIA in December 2023, still in progress.
__________________________________________________________
3) What kind of questions or issues does this case raise about the industry’s willingness to deal with the complications of AI?
(Que tipo de questões ou problemas este caso levanta sobre a vontade da indústria de lidar com as complicações da IA?)
The case involving actress Scarlett Johansson and OpenAI raises important questions about the use of artificial intelligence in the entertainment industry, such as ownership of intellectual property rights in AI-created content that resembles a real person; use of celebrity images and voices to “train” artificial intelligence technologies; transparency and consent in AI models; possible means of ensuring that people know when they are interacting with AI systems and that their data is being used in these models; and responsibility of technology companies developing AI models. Therefore, this case lays base the complexity and ethical challenges associated with the use of AI in the entertainment industry, which increasingly makes use of its benefits. The challenge is to find the balance between technological innovation and the guarantee of individual rights.
__________________________________________________________
4) This case highlights what some critics say is a lack of industry oversight in AI and the need for greater protections for creators, what do you think about this? What types of protections should be established?
(Este caso destaca o que alguns críticos dizem ser a falta de supervisão da indústria em IA e a necessidade de maiores proteções para os criadores. O que você acha disso? Que tipos de proteções devem ser estabelecidas?)
A: In fact, we are still paving the way for the complete regulation of the use of new artificial intelligence technologies so that there is a balance between innovation through safe and reliable systems and the guarantee of rights, including copyright. In Brazil, Bill No. 2338/23 is being presented to the Senate as a substitute for previous projects on the subject and based on a final report by a commission of jurists established for this purpose, in 2022, which, in the phase of public inquiries, received important contributions from the Brazilian Copyright Institute. The project establishes a risk-based regulation and a regulatory modeling based on rights, and brings in its wording specific provisions on the protection of copyright in text and data mining activities for the purpose of training and operation of artificial intelligence systems. In order to foster innovation, there is provision for authorization by the competent authority to operate an experimental regulatory environment for innovation in artificial intelligence (regulatory sandbox) in addition to civil liability rules and code of good practices and governance in AI. Therefore, I believe that the bases of protection for creators were dealt with in our national context, and it is now enough to wait for the regular processing of the bill for regulation to finally happen. Other related points, such as copyright regulation pertaining to products generated through the use of AI systems, depend on the prior regulation of an AI already in operation, that is, they end up being covered by the proposal in progress.
__________________________________________________________
5) What infractions did ChatGPT commit with the voice assistant that imitated Johansson? How can these cases be resolved in Brazilian legislation (if they have laws for these situations)?
(Quais infrações o ChatGPT cometeu com o assistente de voz que imitou Johansson? Como esses casos podem ser resolvidos na legislação brasileira (se há leis para essas situações)?)
A: The use without consent of the voice of actress Scarlet Johansson characterizes a direct violation of a personality right, subject to civil compensation for any nonpecuniary damages arising from this misuse. In the Brazilian legal system and other existing legislative systems, the human voice finds protection in the rights of personality, either as an autonomous right or even as an integral part of the right to image or the right to personal identity. Such rights are directly related to intellectual property rights, among them the so-called copyrights, which also find legal protection, in Brazil, by the Copyright Act (Law 9610/1998). In this case, it is important to highlight the possibility of assigning these copyrights on interpretations and image rights to audiovisual producers, platforms, studios, and channels, which would then become the holders of the right of action in the case of violation.
__________________________________________________________
6) Do you think it is necessary to review copyright or image laws in light of situations like this? If so, what revisions and improvements should be made?
(Você acha que é necessário rever as leis de direitos autorais ou de imagem diante de situações como essa? Em caso afirmativo, que revisões e melhorias deveriam ser feitas?)
A: The current context of insertion of new technologies in society, especially those that rely on artificial intelligence applications, brings with it the need for adjustments in the Copyright Act in force today. This is a fact that cannot be overlooked. Among these adjustments, I suspect the creation and adjustment of limitations and exceptions to copyright, such as the non-violation of these rights in the exercise of text and data mining activities for the purpose of training and operation of artificial intelligence systems (already set out in Bill No. 2338/23) and the definition of possible protection of works created by Artificial Intelligence, so that protection of rights and innovation go hand in hand.
Autor: Raissa Varrasquim Pavon • email: raissa.pavon@ernestoborges.com.br