EFF Warns Courts Against Expanding Copyright Law for AI

Courts face pressure to rewrite copyright principles

The Electronic Frontier Foundation is urging courts to resist calls from rightsholders to dramatically expand copyright protections in response to generative AI, warning that such moves would repeat mistakes from past technology panics.

In amicus briefs filed in cases including Concord Music Group v. Anthropic and In re Mosaic LLM Litigation, EFF argues that copyright owners are asking courts to abandon 300-year-old legal principles based on speculation about AI's impact on creative markets. The digital rights organization points to historical parallels: in the 1980s, Hollywood compared videotape recorders to the "Boston strangler," while composer John Philip Sousa claimed in 1906 that player pianos and gramophones would destroy music composition.

None of those fears materialized. Instead, technologies like cameras sparked new creative movements, including photojournalism and a resurgence in portraiture.

The 'market dilution' theory under scrutiny

Copyright owners in current AI litigation are advancing a "market dilution" theory—arguing that building generative AI tools cannot be fair use because those tools might encourage proliferation of competing works. Under this framework, rightsholders claim they need power to control non-infringing works created by others to prevent an imagined flood of AI-generated content from devastating creative markets.

EFF contends this approach fundamentally misunderstands copyright's constitutional purpose: promoting creation of expressive works for public benefit. Copyright law punishes infringement, not competition. Accepting the market dilution theory would give publishers veto power over any expression that might conceivably compete with works they own, including control over tropes, genres, and styles.

Research challenges displacement claims

The organization also challenges two key assumptions in these cases. First, research indicates that large generative AI models are unlikely to produce infringing works because the more training data a model uses, the less any individual example influences particular outputs.

Second, AI tools aren't necessarily displacing human creativity. EFF cites multiple examples: Boston artist Nettrice Gaskins creates Afrofuturist art including an Octavia Butler portrait displayed at San Francisco Airport; Philadelphia-based Alex Smith reimagines Afrofuturism with queer, plus-sized Black superheroes; and MIT architecture professor Ana Miljački created a documentary film on Yugoslav World War II memorials using generative AI.

AI company Bronze works with musicians like Disclosure and Jai Paul to create songs that never sound identical when played twice, challenging traditional conceptions of music.

Why it matters

The outcome of these cases will determine whether copyright law evolves to address legitimate concerns about AI training or instead becomes a tool for incumbent rightsholders to suppress technological competition. Expanding copyright to control non-infringing derivative works would fundamentally alter the balance between creator rights and public access that has defined copyright for centuries—potentially stifling exactly the kind of creative innovation copyright was designed to encourage.

These details were first reported by the Electronic Frontier Foundation.