Google Bard competitor ChatGPT also getting attention from plaintiffs and regulators.
Allegations have lingered for centuries in far-off corners of the literary world that legendary poet and playwright William Shakespeare, the Bard of Avon, stole some of his most famous works. Broad rejection of such claims has been around as long as they have. Now, another Bard has been accused of stealing the writings of others.
Google Bard — Google’s artificial intelligence chatbot — is trained on misappropriated copyrighted data and private information from services including Gmail and Google Search, plaintiffs charge in a suit filed in U.S. District Court for the Northern District of California (J.L., et al. v. Alphabet Inc., Google Deepmind, and Google LLC, No. 3:23-cv-3440, N.D. Calif.).
Google’s general counsel has been quoted in news reports saying the suit is baseless. The company has always been clear and transparent about its use of public information to train its AI services, he said.
Bard vs. ChatGPT
Google’s answer to the popular ChatGPT service, Google Bard pulls real-time information from websites, databases, and other works, whereas, when asked, ChatGPT itself says its responses are “based on the text that I was trained on up until my knowledge cutoff date in September 2021.” The chatbot adds, “I don’t have real-time access to external information …” Bard also notes the sources it draws its answers from, while ChatGPT has been known to make things up, including case law.
Similar cases have been brought against ChatGPT owner OpenAI (see further down), which reportedly is also being investigated by the FTC for possible unfair and deceptive practices. The Washington Post wrote about a 20-page “demand for records” from the agency which wants to know how OpenAI addresses risks related to its AI models. FTC Chair Lina M. Kahn said in an April 25 statement that federal agencies are watching how AI may be used for nefarious and illegal purposes. “We already see how AI tools can turbocharge fraud and automate discrimination, and we won’t hesitate to use the full scope of our legal authorities to protect Americans from these threats,” said Chair Khan. “Technological advances can deliver critical innovation — but claims of innovation must not be cover for lawbreaking. There is no AI exemption to the laws on the books, and the FTC will vigorously enforce the law to combat unfair or deceptive practices or unfair methods of competition.” Kahn made her comments in a statement jointly released by the FTC, the Civil Rights Division of the U.S. Department of Justice, the Consumer Financial Protection Bureau, and the U.S. Equal Employment Opportunity Commission.
Adults and Children in Proposed Class
Adult and minor plaintiffs in the proposed Google Bard class action are identified only by their initials. One of the adult plaintiffs is described as a best-selling author and journalist. The other is described as a professor and actor who is active on social media. Others include minors who play games online and connect with family and friends via Google Hangouts and other platforms. Different plaintiffs use Gmail, Google’s search engine, YouTube, Google Play, Google Bard, online dating services, and various social media platforms, like Facebook, Snapchat, Instagram, and X (formerly Twitter).
For years, the plaintiffs claim, Google DeepMind and its parent company, Alphabet, have been secretly harvesting the data of hundreds of millions of Americans from sources that include social media posts, subscription-based services, and pirate book publishers in order to train their commercial artificial intelligence products, including Google Bard.
The range of data at issue in the suit is expansive. Personal and professional information, creative and copyrighted works, including audio, video and written material, photographs, and even personal email are allegedly included. This is virtually the entirety of the digital footprint of proposed class members, the complaint says. The companies could have obtained this information legally, plaintiffs say, but opted to scrape it from unassuming users without consent or compensation. Google’s action is allegedly contrary to its own warnings of AI risks and those of the Federal Trade Commission.
Google’s Privacy Policy Change
The complaint maintains that Google quietly updated its privacy policy to reassert that everything shared on the internet is fair game to be used for commercial purposes. Huge collections of public, published, and private data are useful in training AI-boosted “large language models,” or LLMs, to communicate like humans. Especially useful are people’s conversations. In addition to misappropriating data from subscription services and pirated books, Google allegedly scrapes data from its own platforms, like Gmail and Google Search.
Further, since Google’s training data comes from a wide range of news and media sources – some of which lack credibility – the spread of misinformation is more likely, the suit maintains.
Among the allegations is that Google is committing copyright infringement through its generative AI text-to-music capabilities. This appears to be the first such claim.
As for causes of action, the suit says Alphabet and the other defendants have violated California’s Unfair Competition Law and the Digital Millennium Copyright Act. Additional counts include negligence, invasion of privacy, intrusion upon seclusion, conversion, larceny/receipt of stolen property, and unjust enrichment. On behalf of other similar individuals, the plaintiffs seek injunctive relief in the form of a temporary freeze on Google’s commercial development and use of its products until it has implemented practices to honor copyrights and privacy. They also seek damages, restitution, and disgorgement.
ChatGPT Also Draws Suits
The Google Bard suit joins at least two other class actions filed this summer alleging illegal use of copyrighted materials to train bots.
Best-selling authors Paul Tremblay and Mona Awad filed a similar proposed class action in the Northern District against ChatGPT company OpenAI, Inc., on June 28, 2023 (Tremblay, et al. v. OpenAI, Inc., et al., No. 3:23-cv-03223, N.D. Calif.). Tremblay authored “The Cabin at the End of the World.” Awad’s works include “13 Ways of Looking at a Fat Girl” and “Bunny.” According to the suit: “When ChatGPT was prompted to summarize books written by each of the Plaintiffs, it generated very accurate summaries. … The summaries get some details wrong. These details are highlighted in the summaries. This is expected, since a large language model mixes together expressive material derived from many sources. Still, the rest of the summaries are accurate, which means that ChatGPT retains knowledge of particular works in the training dataset and is able to output similar textual content. At no point did ChatGPT reproduce any of the copyright management information Plaintiffs included with their published works.”
Comedian and author Sarah Silverman, along with “Ararat” author Christopher Golden, and “Sandman Slim” author Richard Kadrey, sued OpenAI for copyright violations on July 7, 2023. Silverman, who claims OpenAI violated the copyright of her memoir, titled “The Bedwetter,” says in the proposed class action that the company likely used a pirated version of the work. Others speculate a synopsis of the memoir could have been pulled from descriptions and reviews on the web (Silverman, et al. v. OpenAI, et al.,, No. 3:23-cv-03415-AMO, N.D. Calif.) Comparisons are being made to lawsuits brought several years ago against another Google service. Authors’ claims of infringement against Google Books were rejected, resulting in a 2015 opinion from the Second Circuit in Google’s favor. The Supreme Court declined to hear the case. Authors Guild v. Google 721 F.3d 132 (2nd Cir. 2015)
A highly publicized example of the tension between antitrust and competition, specifically in the labor market, is the ongoing strike between members of the Writers Guild of America and the Alliance of Motion Picture and Television Producers. One aspect of the dispute is the use of AI to create or rewrite scripts. Writers don’t want their original works to be used to train writing bots, either. In general, the WGA issued a report which they say details how “deregulation and the growing dominance of streaming have laid the groundwork for a media landscape where three companies—Disney, Amazon, and Netflix—are poised to be the new gatekeepers of media.” Increased use of AI-generated content, and how to regulate that use, is reportedly a key sticking point in the negotiations.