A war has erupted between the technology and media sectors over generative artificial intelligence. The technology raises profound questions about the ways in which things, such as text, images, and music are produced and used. Legal challenges this year have been surprisingly few and far between, including some from novelists, journalists, and comedians. Who filed lawsuits for copyright infringement.
Noting that their work was used to train large linguistic models, while “Getty Images” launched an attack on “Stability” of artificial intelligence; Because of the use of its photo library, while a lawsuit was filed against “Anthropic” because of the lyrics of the songs.
However, the majority of rights holders were reluctant to move further; Hoping to find ways to share in the spoils of new technology, rather than seeking to thwart and obstruct it. The only two notable cases of agreement between the worlds of technology and media were the Associated Press allowing its archives to be used to train Open AI models.
The second was Axel Springer, owner of Politico, the German newspaper Die Welt, and Business Insider, reaching a broader agreement with the same company early this month.
Thus, the lawsuit filed by “The New York Times” against both “Open AI” and “Microsoft” is a harbinger of what 2024 may hold in store, as “The New York Times” announced the failure of months of negotiations to come up with terms that protect the company’s rights and provide fair compensation.
The early cases that formed the legal basis for search engines are echoed in litigation around generative artificial intelligence. At the time, US courts ruled that using indicators of copyrighted content when creating new “conversional” search engines was considered “fair use.”
The courts found that short excerpts and “thumbnail” images do not represent a substitute for the original content, which put an end to the harm that search engines can inflict on the business of media companies, but this time it carries with it some important differences. In this case, “The New York Times” showed how it helped “Chat” GBT, which is owned by OpenAI, and Bing, the artificial intelligence-enhanced search engine owned by Microsoft, provide verbatim and extended quotations from the reports they produce.
Moreover, while search engines are designed to refer web traffic to other sites, generative AI services, such as ChatGPT, answer questions directly, making them a clear alternative to the original source of the content. These greater legal risks would make AI companies reluctant to test their “fair use” defense before a jury.
However, there are weighty considerations in the opposite direction, beginning with the risk that the jury’s verdict will be a double-edged sword. OpenAI will be able to point out that the company can easily prevent news publishers from using its content if it does not want to use it to train its large language models, a step taken by several publishers, including the New York Times.
Generative AI also threatens to “commodify” several types of information. Once it trains its model on the content it receives from AP and Axel Springer, OpenAI will have less need for more archival news material. This would significantly limit the compensation each publisher would be able to negotiate. , as well as the number of bilateral agreements that artificial intelligence companies will be prepared to reach.
All of this makes a return to the negotiating table the most likely outcome, before any confrontation in the courts. Overall, generative AI promises to create large new markets for media content, but the big question remains: How will the spoils be divided?
Media companies hope to reap value directly from the technology, by training artificial intelligence models on their archives and summarizing their news content, but judging by the wide audience that “ChatGBT” achieved in its first months, smart chatbots and other services appear to be supported by… AI is on its way to being a huge media site itself.
Axel Springer is expected to earn tens of millions of euros annually from its agreement with OpenAI, but this may not be much for a transformative technology that would turn things upside down in the media field, and Axel Springer’s reception Payments of 40 million euros will only add approximately 1% of its annual revenues.
In contrast, news groups face the risk of handing over their audiences to artificial intelligence companies, and the values of their brands may decline if “GBT chat” and its successor become the new sources of knowledge and inspiration of the Internet.
It is impossible to imagine exactly what new services will emerge from generative intelligence, nor even how much value they will have, since it is still in its early stages, and this makes it difficult, more than anything else, for media companies to reach agreement on the terms of their surrender of their future rights. But the pressure on it to reach an agreement will only increase.
2023-12-30 22:03:41
#media #technology #industry #igniting #war #generative #intelligence
