Google has sued three scammers for offering a fake download of its Bard AI chatbot that contained malware capable of stealing credentials for small business' social media accounts.

The web giant's lawsuit [PDF], filed on Monday in a San Jose federal district court against defendants listed as "DOES 1-3", alleges the trio launched social media pages and posts containing links to download of Google’s AI chatbot Bard.

However, the download actually contained not Bard but malware designed to siphon users’ online login details, the court documents state. Google's legal team accused the miscreants of targeting small businesses and advertiser accounts to try and steal their financial information.

"As public excitement in new generative AI tools has increased, scammers are increasingly taking advantage of unsuspecting users," Google's General Counsel Halimah DeLaine Prado wrote.

"We are seeking an order to stop the scammers from setting up domains like these and allow us to have them disabled with US domain registrars. If this is successful, it will serve as a deterrent and provide a clear mechanism for preventing similar scams in the future," she added.

Google is suing the defendants in the US for trademark infringement because the accused used its logos in their fake ads. The Android giant has also alleged violations of unfair competition and contract law, and asked judges to issue a permanent injunction to restrain the trio from offering the fake Bard download.

The web giant also wants damages, including all the money made from the scam. Prado said that the tech titan has filed around 300 takedowns related to this group so far.

Copyright woes also lead to court

In a separate lawsuit, Google also went after another group of miscreants for allegedly abusing the Digital Millennium Copyright Act (DMCA).

The group, consisting of Nguyen Van Duc, Pham Van Thien, plus "DOES 1-20", was accused of making up thousands of fake Google accounts and submitting "thousands of bogus copyright claims" against rival businesses that led to the removal of their websites from Google Search.

"Defendants have weaponized copyright law's notice-and-takedown process and used it not for its intended purpose of expeditiously removing infringing content, but instead to have the legitimate content of their competitors removed based on false allegations," state court documents [PDF]. "Defendants' illegal, fraudulent scheme harms consumers, third-party businesses, and Google; stifles competition; and threatens to tarnish Google's trusted brand."

Speaking of AI and copyright, a federal judge dismissed some of the claims in a lawsuit spearheaded by novelists Richard Kadrey, Christopher Golden, and comedian Sarah Silverman against Meta. At a hearing last week, Judge Vince Chhabria said he struggled to understand the plaintiffs' complaint accusing Meta's Llama language model of violating copyrighted works.

Although the system was trained on works by Kadrey, Golden, Silverman, and others, Judge Chhabria doubted whether the text it generated could be considered copies of their writings.

Chhabria said their arguments "would have to mean that if you put the Llama language model next to Sarah Silverman's book, you would say they're similar," according to Reuters. "That makes my head explode when I try to understand that," he added. Chhabria, however, has given the plaintiffs a chance to amend their complaint against the social media giant.

The group filed a similar case against OpenAI too, which is scheduled to be heard next month.

The Register has asked Meta and the plaintiffs' lawyers for comment. ®

https://www.theregister.com//2023/11/14/google_sues_ai_scammer/