Authors may be eligible for up to $4,500 per published work, though publishers are also expected to receive a share of the compensation.
The legal dispute arose after the works of both global and Australian authors were scraped from open-source databases without permission. This led to claims of intellectual property theft tied to the training of artificial intelligence tools. Anthropic, a US-based tech company founded in 2021, is one of the first AI firms to reach a settlement with affected authors.
The agreement establishes a $2.27 billion (US$1.5 billion) fund for compensation, with estimated payouts of about $3,000 per book before fees. Although payouts may differ based on the author’s portfolio, those with multiple titles listed in the "settlement database" could receive significant financial compensation. The court-approved deal also requires Anthropic to delete all books obtained from LibGen and similar datasets and to destroy AI training materials created from those books.
Well-known Australian titles allegedly used without consent include bestselling works from major authors and acclaimed historians, some of whom have been active in recent copyright debates. In cases where authors are deceased, their estates may still qualify for compensation under this agreement.
The broader outcome of this case could extend well beyond financial payment. It may set a legal and ethical precedent for how AI firms gather training material, especially as issues surrounding data transparency, user consent and ownership of creative work continue to grow. Although this is among the first settlements of its kind, it likely signals the beginning of a wave of legal actions in the fast-changing space of AI and copyright.
