Gharavi v. Google LLC
U.S. District Court, Northern District of California · N.D. Cal. · California bar guidance
Verified September 29, 2026
- Citation
- Gharavi v. Google LLC, No. 25-mc-80164-WHO, 2026 WL 82256 (N.D. Cal. Jan. 12, 2026)
- Decided
- January 12, 2026
Summary
In a DMCA subpoena dispute, Perkins Coie attorney Torryn T. Rodgers filed a declaration for Google stating, "based on a review of Bloomberg Law's search results," that petitioner Nima Gharavi appeared to have represented himself pro se in fifteen cases, including "County of Racine vs. Nima Gharavi, No. 2018TR008316." The court found that while "Bloomberg indicated Gharavi was a party to the County of Racine case, this ended up being untrue." Gharavi, who said he had never litigated in Wisconsin, moved for $5,625 in compensatory sanctions under the court's inherent authority. Google filed a corrected declaration removing the entry.
- AI tool:
- Not mentioned in the order (the error came from Bloomberg Law search results)
How did the court address the AI-related conduct?
Sanctions DENIED. Judge William H. Orrick found the error "neither reckless nor done in bad faith": Rodgers "relied on Bloomberg Law, a well-trusted database and legal search tool, and had no reason to believe that the information provided was false," and not using "a second, outside source" was not reckless. The court called it "a harmless error that does not impact the outcome of the parties' motions." Gharavi's motion to compel and motion to strike were denied, and Google's motion to strike was granted.
Why does Gharavi v. Google LLC matter for law firms using AI?
The order in Gharavi does not mention AI. The case sits in AI hallucination trackers because a sworn declaration listed a case that a research database wrongly attributed to the opposing party. Judge Orrick denied sanctions under the Ninth Circuit’s inherent-authority standard, which requires recklessness plus an improper purpose, finding that reliance on “a well-trusted database and legal search tool” was not reckless and that the error was “harmless.” The order is a useful contrast for firms: in assessing inherent-authority sanctions, this court looked at the source relied on and whether the error mattered, as well as at the error itself.
Sources
Primary sources
- AI attribution: the order does not mention artificial intelligence. The erroneous entry came from Bloomberg Law search results run by the firm's research staff. The entry is included because third-party AI hallucination trackers list it.