A series of high-profile lawsuits in a New York federal court has ignited a legal battle that will determine the future of artificial intelligence (AI) products such as ChatGPT. These AI chatbots, developed by OpenAI and Microsoft, have come under scrutiny for potentially violating copyright and fair competition laws. The lawsuits, filed by renowned authors, media outlets, and The New York Times, argue that OpenAI has built its products on the intellectual property of others without permission.

The legal battle is expected to be challenging for professional writers and media outlets, as copyright attorney Ashima Aggarwal, formerly of John Wiley & Sons, believes they face an uphill battle in court. The lawsuits, which include claims from The New York Times, a group of well-known novelists (including John Grisham, Jodi Picoult, and George R.R. Martin), and bestselling nonfiction writers, allege different violations but all revolve around OpenAI's use of copyrighted works.

Attorney Justin Nelson, representing the nonfiction writers and The Times, argues that OpenAI is essentially using other people's intellectual property without consequences. He criticizes OpenAI's claim that it has the right to use any intellectual property found on the internet since its inception. The Times' lawsuit, filed in December, contends that ChatGPT and Microsoft's Copilot compete with the outlets they are trained on, diverting web traffic and damaging the reputation of copyright holders who rely on advertising revenue.

U.S. District Judge Sidney H. Stein is presiding over all three cases, as well as a fourth lawsuit filed by two nonfiction authors. OpenAI and Microsoft have yet to file formal counter-arguments in the New York cases. However, OpenAI released a public statement dismissing The Times' lawsuit as "without merit" and characterizing the chatbot's ability to reproduce articles verbatim as a "rare bug." The company argues that training AI models using publicly available internet materials falls under fair use, pointing to licensing agreements with organizations like The Associated Press and Axel Springer.

OpenAI's blog post this week downplayed the significance of news content for AI training, stating that large language models learn from an extensive collection of human knowledge and that no single data source, including The New York Times, is crucial for the model's intended learning. The company emphasized its efforts to support a healthy news ecosystem, citing licensing agreements and payments made to access AP's archive of news stories.

The outcome of this legal battle is uncertain. The AI industry's argument relies on the fair use doctrine of U.S. copyright law, which permits limited use of copyrighted materials for purposes such as teaching and research. In response, The Times' legal team asserts that OpenAI and Microsoft's actions do not qualify as fair use, as they exploit the newspaper's investment in journalism to develop substitute products without permission or payment.

Historically, courts have generally sided with tech companies when it comes to copyright infringement claims related to AI systems. For instance, a federal judge in San Francisco dismissed a significant lawsuit against AI image-generators, and another California judge rejected comedian Sarah Silverman's claims that Meta (formerly Facebook) infringed on the text of her memoir. However, the recent lawsuits present more detailed evidence of alleged harms. Copyright attorney Aggarwal notes that courts are generally hesitant to consider the use of copyrighted content for training AI systems as copyright infringement when only a small portion is delivered to users.

Tech companies often cite Google's legal victory in defending its online book library as precedent. In 2016, the Supreme Court upheld lower court rulings that rejected authors' claims of copyright infringement against Google's digitization of millions of books. However, fair use arguments are determined on a case-by-case basis, taking into account economic impact and other factors. Cathy Wolfe, an executive at Wolters Kluwer and a board member of the Copyright Clearance Center, emphasizes that just because something is freely available on the internet does not mean it can be copied or used for commercial purposes.

As the lawsuits proceed, the legal battle between OpenAI, Microsoft, and the authors and media outlets will shape the future of AI products and their relationship with copyrighted works. The outcome will have far-reaching implications for the AI industry and the boundaries of fair use in the digital age.