The company strongly defends the honor of ChatGPT after allegations of copyright infringement.
Microsoft strongly countered The New York Times accusations that Windows and OpenAI infringe copyright by using the publication's articles to teach language models like ChatGPT.
In court documents, Microsoft representatives drew a parallel with the situation in the early 1980s, when the Film Distributors ' Association tried to restrict the distribution of video recorders. The NYT's claims against OpenAI are ostensibly aimed at thwarting a "new technological breakthrough."
The documents claim that the newspaper's claims that GPT-based products "harm The New York Times" and "pose a deadly threat to independent journalism "are nothing more than fantasies about an"apocalyptic future."
The NYT lawsuit is one of many that OpenAI has recently faced in connection with controversial methods for teaching large language models (LLMs). The publication claims that significant amounts of its original content were used to train AI without permission, and gives examples that, in its opinion, prove that ChatGPT was trained on illegally copied materials.
Microsoft's defense team does not appear to deny the use of journalistic content, but states: "Contrary to The Times' claims, copyright law is not a big obstacle to LLM, as it once was with video recorders, pianols, copiers, PCs, the Internet or search engines."
The corporation believes that the NYT's arguments about copyright infringement are greatly exaggerated, because other technological innovations that have now become commonplace were also criticized in the same way.
According to Microsoft, the methods by which the NYT decided to prove the fact of a violation do not fully reflect how the GPT tools are actually used. "The Times created unrealistic queries by deliberately forcing AI to produce snippets of text that match its content," the statement said.
Microsoft wants to dismiss three specific NYT lawsuits. One of the arguments is that the company should not be held responsible for potential copyright violations by end users.
Lawyers say that in order for GPT tools to produce verbatim quotes from The New York Times materials, the user must know in advance where these quotes came from. The documents note :" In addition, the texts cited in the lawsuit cannot be considered complete copies of works at all, they are just small fragments."
Earlier, OpenAI accused The New York Times that someone commissioned by the publication "hacked" ChatGPT and forced it to reproduce verbatim quotes from publications. At the same time, as explained in OpenAI, "hacking" meant using a regular system with provocative requests.
However, NYT legal adviser Ian Crosby denied these allegations, saying that they were simply testing OpenAI products for copyright compliance.
Microsoft strongly countered The New York Times accusations that Windows and OpenAI infringe copyright by using the publication's articles to teach language models like ChatGPT.
In court documents, Microsoft representatives drew a parallel with the situation in the early 1980s, when the Film Distributors ' Association tried to restrict the distribution of video recorders. The NYT's claims against OpenAI are ostensibly aimed at thwarting a "new technological breakthrough."
The documents claim that the newspaper's claims that GPT-based products "harm The New York Times" and "pose a deadly threat to independent journalism "are nothing more than fantasies about an"apocalyptic future."
The NYT lawsuit is one of many that OpenAI has recently faced in connection with controversial methods for teaching large language models (LLMs). The publication claims that significant amounts of its original content were used to train AI without permission, and gives examples that, in its opinion, prove that ChatGPT was trained on illegally copied materials.
Microsoft's defense team does not appear to deny the use of journalistic content, but states: "Contrary to The Times' claims, copyright law is not a big obstacle to LLM, as it once was with video recorders, pianols, copiers, PCs, the Internet or search engines."
The corporation believes that the NYT's arguments about copyright infringement are greatly exaggerated, because other technological innovations that have now become commonplace were also criticized in the same way.
According to Microsoft, the methods by which the NYT decided to prove the fact of a violation do not fully reflect how the GPT tools are actually used. "The Times created unrealistic queries by deliberately forcing AI to produce snippets of text that match its content," the statement said.
Microsoft wants to dismiss three specific NYT lawsuits. One of the arguments is that the company should not be held responsible for potential copyright violations by end users.
Lawyers say that in order for GPT tools to produce verbatim quotes from The New York Times materials, the user must know in advance where these quotes came from. The documents note :" In addition, the texts cited in the lawsuit cannot be considered complete copies of works at all, they are just small fragments."
Earlier, OpenAI accused The New York Times that someone commissioned by the publication "hacked" ChatGPT and forced it to reproduce verbatim quotes from publications. At the same time, as explained in OpenAI, "hacking" meant using a regular system with provocative requests.
However, NYT legal adviser Ian Crosby denied these allegations, saying that they were simply testing OpenAI products for copyright compliance.
