In response to a lawsuit filed by tHe is the New York Times, news outlets accused OpenAI of using news content to train AI models, and OpenAI brought the receipts. Leading AI developers drew on their oft-stated commitment to the news industry, declaring: “We support journalism, work with news organizations, and new york times “The lawsuit is without merit,” he said.
OpenAI was also criticized. new york times “About incomplete reporting” new york times The company suggested that the examples used in the newspaper were from older articles widely available on third-party websites, and also implied that: new york time We created AI prompts to generate the most damning evidence.
“It appears that we intentionally manipulated the prompts to include long article excerpts to force our models to regurgitate,” OpenAI said. new york times Acted in bad faith by providing unnatural prompts as evidence.
“Even with these prompts, our models typically don’t behave like this: new york times This suggests that the model was instructed to regurgitate or that cases were selected over multiple trials.”
Prompt manipulation is a common practice where people use specific prompts to trigger very specific responses that cannot be obtained under normal conditions to trick AI models into performing tasks they were not programmed to do.
OpenAI emphasized collaboration with the news industry.
“We are working hard on our technology design process to support news organizations,” the company wrote, highlighting its goals for the deployment of AI tools to support reporters and editors and the parallel growth of both AI and journalism. OpenAI recently partnered with publisher Axel Springer. rolling stone-Provides more accurate news summaries.
To solve the problem of content ‘reflux’ new york times OpenAI acknowledges that this is a rare but existing problem that it is working to mitigate.
“Memorization is a rare failure of our continuously evolving learning process,” they explained and defended their training methods. “Training an AI model using publicly available Internet material is fair use.”
Nonetheless, OpenAI has acknowledged the validity of ethical considerations by providing an opt-out process for publishers.
AI training and content storage
The battle between content creators and AI companies appears to be a zero-sum game for now. Because at the root of it all is the fundamental way we train AI models.
These models are developed using massive data sets consisting of text from a variety of sources, including books, websites, and articles. Other models use pictures, illustrations, movies, voices, and songs, depending on what they were trained to do. However, these models do not retain specific articles or data. Instead, it analyzes this material to learn language patterns and structures.
This process is important to understand the nature of the claims, OpenAI’s defense, and why AI trainers believe their business is using the content in a fair way. This is similar to how art students study other artists or art styles to understand their characteristics.
But creators new york times and best-selling authors claim companies like OpenAI are using their content maliciously. They claim their intellectual property is being exploited without permission or compensation, potentially leading to AI-generated products that compete with and divert audiences from the original content.
that much new york times OpenAI sued OpenAI, emphasizing that the use of its content without explicit permission undermines the value of original journalism and could have a negative impact on the production of independent journalism and its costs to society. And no matter how sophisticated the prompt is, you could argue that if it “regurgitates” any kind of copyrighted content, it’s because it was used.
It is up to the court to decide whether it was used fairly or unfairly.
This legal battle is part of a legal movement that could shape the future of AI, copyright law, and journalism. As events unfold, they will undoubtedly impact the conversation surrounding the integration of AI into content creation and the rights of intellectual property owners in the digital age.
Nonetheless, OpenAI does not believe this is a zero-sum scenario. Despite criticizing the lawsuit’s core content, Altman’s company said it was prepared to extend an olive branch and look for a positive outcome somewhere.
“We hope for a constructive partnership that will: new york times “We respect our long history, including reporting on the first working neural network 60 years ago and defending First Amendment freedoms.”
Edited by Ryan Ozawa.