Authors, publishers sue Google over alleged AI copyright infringement

Hachette Book Group, Cengage Learning, Elsevier, and writer Scott Turow person filed a suit against Google successful a national United States tribunal successful New York, alleging the Silicon Valley tech elephantine committed copyright infringement portion grooming its Gemini AI models.

“Google willfully sidestepped this longstanding strategy designed to support copyrights and compensate authors and publishers done a bid of deliberate choices to make Gemini1,” the astir 60-page ailment filed connected Friday said.

Recommended Stories

• list 1 of 4US user prices driblet successful June arsenic vigor costs tumble
• list 2 of 4New York imposes landmark one-year prohibition connected ample information centres
• list 3 of 4My Twitter, not X
• list 4 of 4How US-Iran escalation volition trial Iraq’s balancing act

The lawsuit alleges that Google archetypal copied books arsenic root worldly done Google Books, claiming that the institution utilized books it “obtained for strictly constricted purposes successful transportation with Google Books and different Google services”. It besides alleges that Google “downloaded web scrapes of virtually the full internet, including from known pirate sources and from down morganatic paywalls”.

It further alleges that Google copied those works without support to bid its AI models and continues to bash so, contempt those uses allegedly falling extracurricular the scope of existing agreements.

The suit claims the institution was afloat alert of the ineligible risks, alleging that interior documents warned utilizing books to bid AI models was “highly problematic for Google,” and could pb to arsenic overmuch arsenic $100bn successful fines.

“At nary constituent did Google pass authors and publishers that Google was copying their works arsenic root worldly to make and bid AI models,” the suit alleges.

“The idea, successful short, is that immoderate just usage [a ineligible doctrine that allows constricted usage of copyrighted worldly without support for purposes similar reporting, education, and probe successful a constricted scope] statement that Gemini has would arguably beryllium mooted by the information that they allegedly acquired the books unlawfully,” Kirk Sigmon, founding spouse who looks astatine exertion and IP instrumentality astatine KellDann Law, told Al Jazeera. “It’s an absorbing contented that has a batch of analyzable dimensions, successful nary tiny portion due to the fact that it tin beryllium hard to beryllium what was oregon wasn’t successful a grooming corpus.”

The suit follows a erstwhile effort by Hachette Book Group and Cengage successful February to articulation a pre-existing people enactment suit primitively brought by a radical of authors successful 2023.

“The scope of the ailment underscores that authors and publishers are agreed successful the extremity of protecting their invaluable intelligence spot rights successful works of fiction, nonfiction, children’s books, memoirs, and poetry, arsenic good arsenic acquisition works and scholarly articles that span thousands of taxable areas and probe developments,” Hachette said successful a connection pursuing the complaint.

Google did not respond to Al Jazeera’s petition for comment.

A question of lawsuits

It is acold from the archetypal suit brought against an AI institution by authors and publication publishers implicit alleged copyright infringement.

There is besides a pending suit against OpenAI brought by authors including George RR Martin, writer of Game of Thrones, and the Authors Guild. In October, a national justice denied OpenAI’s effort to disregard the case.

However, a antithetic suit brought by a radical of authors against Meta did not spell successful the authors’ favour. In 2025, a radical of authors led by Richard Kadrey alleged that Facebook and Instagram’s genitor institution utilized copyrighted books to bid its AI models. A national justice ruled that the AI grooming met the ineligible requirements for “fair use”.

“They [the copyright lawsuits] each typically travel the aforesaid roadworthy map,” Michael Goodyear, subordinate prof astatine New York Law School, told Al Jazeera.

“The basal claims are you took copyrighted works and utilized them for training. These were unlawful copies, that’s the grooming argument. Some besides much explicitly marque arguments astir infringing output.”

But Oli Huggins, CEO of ExpertEdge and VP of Partnerships astatine Packt Publishing — a steadfast whose titles person been utilized successful an Anthropic piracy lawsuit and who told Al Jazeera helium has been approached by AI companies seeking to wage to bid connected the company’s information — said the situation is that erstwhile accusation is utilized for training, it is hard to beryllium whether an output constitutes copyright infringement.

“Proving what happened wrong a exemplary is different cardinal difficulty. Once the ovum is baked into the cake, it is highly hard to place it, quantify its publication oregon beryllium precisely which transcript of a publication was used. A exemplary whitethorn uncover familiarity with a enactment without reproducing capable verbatim substance to found the evidentiary way a claimant needs,” Huggins told Al Jazeera.

Huggins said the licensing offers presently being made are not sustainable for publishers.

“The economics stay profoundly unattractive. Offers presently circulating tin worth a perpetual AI-training licence astatine astir $10 per title.”

There is simply a question of lawsuits crossed content-driven industries, including quality and music, against AI companies implicit alleged copyright infringement arsenic well.

CNN filed a suit against Perplexity, alleging the institution illegally copied much than 17,000 stories to bid its models, which generated contented that was “identical oregon substantially akin to CNN’s content”, according to the ailment filed successful May.

Last week, 17 quality organisations, including The New York Times, accused OpenAI of withholding grounds successful a lawsuit primitively brought by the paper successful 2023, alleging the Sam Altman-led AI institution committed copyright infringement portion grooming ChatGPT.

In the euphony sector, Hagens Berman, a salient people enactment instrumentality firm, filed a people enactment suit against AI euphony generator Suno, alleging it trained its models connected autarkic musicians’ enactment without their consent.

In January, Universal Music Group sued Anthropic, alleging the Dario Amodei-led institution committed copyright infringement by utilizing 20,000 songs to bid its Claude exemplary without permission.

But determination are inactive looming questions astir who is yet liable for reproduced worldly that appears to beryllium copied, according to Goodyear, that the courts person yet to answer.

“If the idiosyncratic is actively trying to get the exemplary to infringe, that could mean the idiosyncratic is yet connected the hook alternatively than the AI system,” Goodyear said. “This is inactive an unfastened question that courts successful the United States haven’t truly grappled with.”