• by Wendy Davis  @wendyndavis, October 17, 2023

Google is asking a federal judge to dismiss a class-action complaint claiming the company engaged in “mass theft of personal information” in order to develop artificial intelligence products such as the chatbot Bard.

In a motion filed Monday, Google describes the complaint as an “anti-AI polemic” that's based on the “false premise” that training generative artificial intelligence on information shared online equates to theft.

“Using publicly available information to learn is not stealing,” Google writes in papers filed with U.S. District Court Judge Araceli Martínez-Olguín in the Northern District of California. “Nor is it an invasion of privacy, conversion, negligence, unfair competition, or copyright infringement.”

The papers come in response to a lawsuit brought in July by eight individuals who claimed Google was “secretly stealing everything ever created and shared on the internet.”

The 90-page complaint didn't name the litigants, but described one as a best-selling author and journalist based in Texas, and another as a professor and actor from California.

The plaintiffs accused Google of harvesting “all our personal and professional information, our creative and copywritten works, our photographs, and even our emails -- virtually the entirety of our digital footprint.”

The suit includes claims that Google infringed copyright, and that it violated users' privacy.

Google says the complaint should be dismissed at an early stage for numerous reasons, including vagueness.

“Next to nothing illuminates the core issues, such as what specific personal information of Plaintiffs was allegedly collected by Google, how (if at all) that personal information appears in the output of Google’s generative AI services, and how (if at all) Plaintiffs have been harmed,” the company writes. “Without those basic details, it is impossible to assess whether Plaintiffs can state any claim and what potential defenses might apply.”

The company adds that even though the complaint references “theft” of personal information, the web users don't identify what data was allegedly misused.

"Plaintiffs have done nothing more than generically allege that Google 'scraped' and exploited 'the entire internet,' including generic, broad categories of 'personal information," Google writes.

"Nowhere do plaintiffs identify the specific information at issue or explain why that information was private and sensitive," Google adds. "Nor do they explain how their privacy was violated; for example, they do not allege that Google accessed private servers or that Bard has ever shared a single piece of personal information about any of them."

Google adds that the complaint doesn't allege facts showing how any particular copyrighted work was infringed.

Many of the complaint's allegations draw on published reports, including an April 19 Washington Post article about the data used to train various chatbots. According to that article, the copyright symbol appeared more than 200 million times in one of the datasets.

The plaintiffs also noted that Google on July 1 updated its privacy policy to say the company uses publicly available data to train artificial intelligence products -- including Bard and Google Translate. The prior policy also allowed Google to use publicly available information to train “language models,” but didn't reference Bard by name.

Google general counsel Halimah DeLaine Prado stated when the lawsuit was filed that the company has been clear about its use of data from public sources to train artificial intelligence models.

"American law supports using public information to create new beneficial uses, and we look forward to refuting these baseless claims," Prado stated.

Google isn't the only company facing litigation over generative artificial intelligence. OpenAI is also facing several copyright infringement lawsuits -- including one brought in federal court in New York by the Authors Guild and novelists including George R.R. Martin, Jodi Picoult and Scott Turow.