Anthropic, an AI startup established in 2021, has secured a momentous settlement of US$1.5 billion (AU$2.28 billion) in a class-action copyright lawsuit. The legal action was initiated in 2024 by novelist Andrea Bartz, along with non-fiction writers Charles Graeber and Kirk Wallace Johnson.
Should the settlement receive the endorsement of the presiding judge, the company will distribute approximately US$3,000 to authors for each of the estimated 500,000 books encompassed in the agreement.
Additionally, it will annihilate illegally downloaded books and refrain from employing pirated works in future chatbot training.
This settlement marks the largest copyright agreement in U.S. history, setting a vital legal benchmark for the evolving dynamics between AI enterprises and their content creators.
Its ramifications will undoubtedly extend to various other ongoing copyright litigations against AI firms, including OpenAI, Microsoft, Google, and most recently, Apple. In June, Meta successfully defended itself in a copyright dispute, although the ruling left avenues open for potential further lawsuits.
The settlement follows a landmark ruling issued in June 2025 concerning AI development and copyright, which delineated a distinction between legal AI training and the illicit acquisition of content. Allegations against Anthropic suggest it unlawfully downloaded over seven million books from two online “shadow libraries” between June 2021 and July 2022.
The parties involved are anticipated to finalize a list of authors whose works will receive compensation by September 15.
Cautious Optimism
In Australia, the response to the potential settlement has been tempered with cautious optimism. Stuart Glover, head of policy at the Australian Publishers Association, commented:
We welcome these court-enforced measures towards accountability, but this settlement underscores the necessity for AI companies to honor copyright and fairly compensate creators—rather than exploiting what they can.
Moreover, for Australian authors and publishers whose creations have been exploited without remuneration, this represents a compelling appeal to the Australian government to uphold copyright and ensure that AI firms remunerate for their usage.
Lucy Hayward, CEO of the Australian Society of Authors, added:
While the complete details remain undisclosed, this settlement is a promising recognition that AI companies cannot simply appropriate authors’ and artists’ works for developing their extensive language models.
Lucy Hayward emphasizes the need for ongoing compensation for Australian authors whose works have been utilized to train AI models.
Australian Society of Authors
Nevertheless, it is vital to note that authors will only receive compensation if their publishers have registered their works with the U.S. copyright office within a specified timeframe. Hayward voiced concerns regarding this stipulation, as many of the seven million allegedly pirated works were penned by international authors, and “we suspect numerous global authors may forfeit settlement compensation.”
She has urged the Australian government to enact new legislation mandating that tech companies provide “continuous compensation to creators for Australian texts used in offshore model training.”
Legal Risks
In June, U.S. judge William Alsup determined that employing books for AI training did not constitute a violation of U.S. copyright law. However, he concluded that Anthropic would still face trial regarding its use of pirated materials to curate its library.
Judge Alsup has since criticized the settlement, citing its shortcomings. He has scheduled a subsequent hearing for September 25, stating, “We’ll see if I can hold my nose and approve it,” he remarked.
Failure to secure settlement approval could expose Anthropic to significantly more severe financial liabilities. A trial is slated for December; if the company loses, U.S. copyright law permits statutory damages of up to $150,000 per infringement for willful violations.
Legal analyst William Long from Wolters Kluwer forecasts that potential damages in a trial could escalate into billions, posing a risk of jeopardizing or even bankrupting the company.
Recently, Anthropic secured a new funding round amounting to US$13 billion, pushing its valuation to $183 billion. Keith Kupferschmid, president and CEO of the Copyright Alliance, argued this demonstrates that “AI firms can afford to compensate copyright holders without impeding their ability to innovate and compete.”
For Mary Rasenberger, CEO of the Authors Guild, the landmark settlement represents “an excellent triumph for authors, publishers, and rights holders.” Rasenberger anticipates that “the settlement will pave the path for more licenses that grant authors both remuneration and autonomy over how their works are utilized by AI companies, as is fitting in a properly functioning free market.”
A Step Forward
Although this particular settlement may provide meager assistance to Australian authors and publishers whose works are not registered with the U.S. Copyright Office, it is undeniably potentially favorable news for global creators. It signifies progress towards the establishment of a structured licensing framework.
Australian copyright law lacks a U.S.-style “fair use” exception, which AI companies argue protects their training methodologies. Calls for legislative adjustments have arisen, as major AI players, including Google and Microsoft, press the Australian government for copyright exemptions.
The recent interim report by the Productivity Commission proposed a text and data mining exception to the Australian Copyright Act, permitting AI training on protected Australian works. This proposal has encountered substantial opposition from the Australian Society of Authors and the publishing sector.
As Arts Minister Tony Burke stated in August 2025, the government harbors “no plans, no intention, and no appetite to weaken” copyright protections.
While the Australian publishing industry does not entirely oppose AI developments, significant legal and ethical challenges persist. The Australian Publishers Association advocates for government policies on AI that prioritize a clear ethical framework, transparency, adequate incentives, and safeguards for creators, aiming for a balanced approach that allows both AI innovation and cultural industries to thrive.
Agata Mrva-Montoya is President of the Round Table on Information Access for People with Print Disabilities.
This article was originally published on The Conversation. Read the original article.
Source link: Inkl.com.
