The debate over whether generative AI companies can use copyrighted material to train their models is increasingly becoming a debate about who gets to use information on the internet. The Electronic Frontier Foundation (EFF) is warning that efforts to narrow copyright’s fair use protections could affect far more than AI developers, potentially creating legal uncertainty for researchers, journalists, artists, and ordinary internet users.

In an article published in October 2026, EFF argues that changing fair use rules to address generative AI could undermine everyday online activities and restrict research that benefits the public. The organization has also filed an amicus brief in In re OpenAI, urging the court to consider the broader consequences of its decision.

The legal dispute comes as publishers, authors, and other copyright holders pursue lawsuits against major AI companies over their use of copyrighted works to train generative AI models. While these cases are often framed as a conflict between creators and powerful technology companies, EFF believes the outcome could reshape copyright protections for everyone.

Why fair use matters beyond AI training

Fair use allows copyrighted material to be used in certain circumstances without obtaining permission from the copyright holder. It supports activities such as research, criticism, reporting, commentary, and transformative creative work.

EFF argues that some of the legal arguments being advanced by copyright plaintiffs could weaken these protections. One particular concern is the suggestion that copyrighted material must be lawfully obtained for a subsequent use to qualify as fair use.

According to EFF, such a requirement could create legal uncertainty around ordinary internet activities. Opening emails, visiting websites, clicking embedded links, sharing pictures, streaming music, and watching videos all involve making temporary or intermediate copies of digital content.

If those copies were subjected to stricter copyright requirements, everyday internet use could become considerably more complicated. Users would potentially face questions about the origin and licensing status of content they access through routine online activities.

Fair use helps prevent the internet from becoming a place where people must individually verify copyright permissions every time they interact with digital material.

Researchers and journalists could face new restrictions

The potential consequences extend to investigative journalism and academic research.

EFF points out that researchers, journalists, and nonprofit organizations frequently use web scraping and crawling to collect publicly available information. These techniques can help identify patterns, investigate corporate behavior, study misinformation, and document harmful online activity.

For example, journalists and nonprofit researchers have used scraping techniques to investigate potentially anticompetitive practices by technology companies and examine misinformation and hateful content published on X.

These methods can resemble the data collection techniques used by generative AI developers. However, the purpose of the research can be entirely different, particularly when it involves investigating powerful companies or serving the public interest.

EFF argues that restricting fair use based on how copyrighted material was initially collected could place these activities at risk. Smaller research organizations and independent journalists may also lack the financial resources to navigate complicated licensing arrangements or defend themselves against copyright claims.

The organization also highlights the implications for academic publishing. Under an increasingly restrictive interpretation of copyright, even literature reviews that summarize and analyze findings from existing research could face additional legal uncertainty.

AI tools also support accessibility and scientific research

EFF acknowledges that creators and workers deserve protection as artificial intelligence changes creative industries. It also argues that technology companies should be held accountable for harmful practices.

However, the organization believes that broadly restricting copyright exceptions is not the right way to address these concerns.

General-purpose AI models can support applications beyond generating text and images. EFF highlights Be My AI, an accessibility tool from Be My Eyes that uses OpenAI’s GPT models to provide visual descriptions for people who are blind or have low vision.

AI-assisted tools have also contributed to scientific research and medical development, including work toward treatments for diseases that were previously considered difficult or impossible to treat.

EFF argues that restricting the use of copyrighted material for transformative research could make it harder to develop such applications.

The broader concern is that copyright rules designed around disputes involving commercial AI companies could unintentionally restrict technologies and research that provide meaningful public benefits.

Mandatory AI licensing could strengthen the biggest technology companies

Another major concern raised by EFF is the proposal to require AI developers to obtain licenses for copyrighted material used during model training.

At first glance, licensing might appear to offer a straightforward way to compensate creators. However, EFF argues that mandatory licensing could create significant financial and administrative barriers for smaller AI developers, independent researchers, and academic institutions.

Large technology companies would be better positioned to negotiate licensing agreements and absorb the associated costs. Smaller competitors could struggle to access the same training material, potentially concentrating AI development among a handful of well-funded companies.

EFF cites law professor Lawrence Lessig, who has warned that requiring licenses for all AI training could make the process affordable only for a small number of technology giants.

The organization also questions whether additional copyright control would necessarily translate into better compensation for individual creators. Publishers, record labels, and other large rights holders often control the commercial rights to creative works, meaning additional licensing revenue would not automatically reach the artists and workers who produced them.

The copyright debate could shape the future of the internet

The disagreement over AI training is about more than whether companies should pay to use copyrighted books, articles, images, and other material. It also raises questions about how information can be accessed, analyzed, transformed, and shared online.

EFF’s position is that copyright should continue serving its broader purpose of encouraging creativity and innovation rather than giving existing rights holders greater control over how information is used.

The organization is not arguing that AI companies should be exempt from accountability. Instead, it wants courts to recognize the distinction between addressing harmful commercial practices and restricting fair use rights that support research, journalism, accessibility, and everyday internet activity.

The outcome of In re OpenAI could therefore have implications well beyond the companies involved in the litigation. For ordinary internet users, the stakes include the freedom to interact with online content without constantly worrying about the copyright status of every digital copy. For researchers and independent creators, they include the ability to build on existing knowledge without facing licensing barriers designed primarily around the resources of major corporations.