Authentic Brands Group, which holds the rights to Elvis Presley’s recordings and likeness, filed a lawsuit on Friday against the artificial-intelligence company Anthropic. The complaint alleges that Anthropic trained its Claude model on 274 Presley tracks without obtaining permission. The suit seeks statutory damages of $150,000 for each infringed song, a total that the filing puts at roughly $41 million (about £31 million).
AI developers have already faced copyright actions from the Recording Industry Association of America and major labels such as Sony, Universal and Warner. However, this filing marks one of the first instances in which an artist’s estate directly targets an AI firm. Legal experts suggest the outcome could shape how future systems are permitted to incorporate existing music catalogues.
The complaint claims that Claude has generated infringing output for virtually every song available online, including Presley classics such as “Can’t Help Falling in Love,” “All Shook Up” and “A Little Less Conversation.” By citing 274 tracks, the plaintiffs calculate damages at $150,000 per song, arriving at an aggregate figure of roughly $41 million.
Lead counsel Richard Busch wrote that the suit aims to enforce exclusive rights against a defendant whose entire business model, he says, relies on copyright infringement at an unprecedented scale. The filing also alleges that Anthropic’s guardrails were ineffective, noting that the lawsuit states Anthropic “knowingly engaged and continues to willfully engage” in copyright infringement.
AI music services operate by ingesting large volumes of recorded tracks to learn patterns of melody, harmony and lyrical structure, then allowing users to generate new compositions via text or voice prompts. Musicians have repeatedly warned that such systems repurpose their work without consent and fear that low-quality AI-generated “slop” could displace authentic human-crafted songs.
Streaming platform Spotify recently reported that it had taken down roughly 50 million tracks it identified as AI-generated slop from its recommendation algorithms over the past year. The move reflects growing industry pressure to curb unlicensed synthetic content and underscores the broader tension between AI innovation and established copyright frameworks.