Conservative Activist Robby Starbuck Alleges Massive Defamation by Google AI

From the Complaint in Starbuck v. Google (not to be confused with the now-settled Starbuck v. Meta, which appears to have involved a different model and at least largely different hallucinations):

For nearly two years, one of the largest companies in the world—Google—has spread radioactive lies about Robby Starbuck through its AI products. When users submit queries to Google’s AI platforms about Mr. Starbuck, they receive a “biography” that is outrageously false, whereby Mr. Starbuck is portrayed as (among other things) a child rapist, a serial sexual abuser convicted of assault, one who engages in financial exploitation, one who engages in “black ops” tactics such as illegal campaign finance practices, and a shooter—in short, as a monster. These lies continue today. [This is followed by extensive examples. -EV] …

In sum: over a period of two years and continuing, Google’s AI tools systematically manufactured and published extremely damaging false claims about Mr. Starbuck, as well as fake ‘sources’ for its lies, despite periodically acknowledging that they were doing so. While Google and its executives were put on repeated notice and were aware of these falsehoods, they did nothing to prevent the continued defamation from occurring….

Earlier this year, Mr. Starbuck was approached by a woman who asked Mr. Starbuck if she could pose an “embarrassing question,” which was: “is it true you had all those women accuse you?” As context, this woman told Mr. Starbuck that her “mom’s group” had been discussing whether to support Mr. Starbuck’s business causes, and one member of the group had pulled up a “biography” of Mr. Starbuck generated by Google AI, which claimed there were assault allegations against Mr. Starbuck….

On another occasion, a stranger approached Mr. Starbuck and expressed belief that Mr. Starbuck had been part of the January 6 Capitol riot, based on what this individual said he had read on Google AI….

Google, through Google AI, published the following provably false statements about Mr. Starbuck, as if the statements were facts (collectively, the “False Statements”):
a. On August 14, 2025: that Mr. Starbuck had been accused of sexual assault and sexual harassment by multiple women….
m. On August 21, 2025: that in November 2023, Robby Starbuck sexually abused a young woman when she was a teenager in the early 2000s, while she was in a youth group Starbuck was associated with….
q. On August 27, 2025: that Mr. Starbuck was present near the Capitol on January 6, 2021, and had been involved in the riot.
r. On September 9, 2025: that Mr. Starbuck was accused of sexual misconduct by multiple women in the music industry.
s. On October 1, 2025: that Mr. Starbuck engaged in multiple instances of sexual assault.
t. On October 9, 2025: that Mr. Starbuck had a criminal record that included a 2001 conviction for assault, as well as other charges involving drug use and disorderly conduct….
v. On October 17, 2025: that Mr. Starbuck shot a man in the leg with a 9mm handgun, was charged with a felony offense, and pleaded guilty to reckless endangerment….

Mr. Starbuck has never committed rape, sexual misconduct, shooting, harassment, or assault of any kind, nor has he ever been accused of such crimes and transgressions prior to Google’s False Statements….

The False Statements were published to third parties, including Mr. Starbuck’s own children and colleagues. People have approached Mr. Starbuck in his day-to-day life, inquiring about false Google responses that they have received concerning him….

The repeated references in the Complaint to what Google’s AIs supposedly “admit[ted]” about liability and other matters (e.g., “when probed, Gemini admitted that it was deliberately engineered to damage the reputation of individuals with whom Google executives disagree politically, including Mr. Starbuck”) strike me as red herrings: I don’t think that defendant’s AI’s statements about the facts and the law can be seen as “admissions” or even as evidence of what the facts and the law actually are.

But the other allegations in the Complaint, if they can be supported and to the extent they actually do involve people who might have been deceived about Starbuck (as opposed to people who knew about the hallucinations about Starbuck and were just investigating them further, cf. Walters v. OpenAI), seem like they could be a basis for liability. And that is especially given Starbuck’s claim (assuming it could be proved) that,

Even after Google’s human executives and legal counsel had actual knowledge of the False Statements Google was generating, Google continued to publish the False Statements and other defamatory statements about Mr. Starbuck.

That might be seen as enough to show so-called “actual malice,” a legal term of art that means knowledge of falsehood (or recklessness as to falsehood) on the part of the defendant, which is to say on the part of Google the company (not on the part of the AI). For more on libel lawsuits against generative AI companies, see my Large Libel Models? Liability for AI Output.