On October 13, 2026, the District Court for the Northern District of California approved a $68 million class action settlement in Kumandan v. Google LLC, No. 19-cv-04286-BLF, 2026 WL 3084349, resolving allegations that Google’s voice-activated devices recorded and disclosed private conversations without users’ consent.
Turns out, “Hey Google” wasn’t necessarily the only thing getting Google’s attention.
The lawsuit stems from allegations that Google’s devices equipped with Google Assistant recorded, transmitted, and processed users’ private conversations without their knowledge or consent. Plaintiffs alleged that, based on Google’s public representations, Terms of Service, and Privacy Policy, users reasonably expected that Google Assistant-enabled devices would only record conversations after receiving a voice command, such as “Hey Google,” or being manually activated.
However, Plaintiffs alleged that the devices could activate under other circumstances, including false activations, referred to as “False Accepts,” and certain alarm events, such as breaking glass or smoke alarms. Plaintiffs further alleged that Google transmitted some of these recordings to unauthorized third-party contractors.
The consolidated litigation began in 2019 and involved several claims, including violations of the federal Wiretap Act, 18 U.S.C. § 2510 et seq., the Stored Communications Act (“SCA”), 18 U.S.C. § 2701 et seq., California Penal Code § 632, intrusion upon seclusion, invasion of privacy, breach of contract, and violations of California’s Unfair Competition Law.
Following four amended complaints and multiple rounds of motions to dismiss, Plaintiffs moved for class certification in July 2022. The Court granted certification of the Purchaser Class, but declined to certify the proposed Privacy and SCA Classes.
In February 2025, Google moved for summary judgment, arguing that the Purchaser Class could not establish most of the essential elements of its claims. Meanwhile, the parties continued settlement negotiations, ultimately reaching an agreement under which Google would pay $68 million to resolve the litigation.
The Settlement
The parties moved for preliminary approval of the settlement, which the Court granted. The settlement established two classes:
Purchaser Settlement Class: All users who purchased a Google-made device between May 18, 2016, and the date of preliminary approval of the settlement.
Privacy Settlement Class: All Google Assistant users and members of their households during the settlement period whose communications were recorded or otherwise obtained as a result of a False Accept, or whose communications were recorded or otherwise obtained and disclosed to a third-party review vendor.
The parties estimated that Purchaser Class members would recover approximately $18 to $56 each, while Privacy Class members would receive approximately $2 to $10 each.
To obtain compensation, class members were required to submit claims online or by mail, provide basic identifying information, and identify the Google account associated with the relevant device. Purchaser Class members were also required to provide device information, serial numbers, and proof of purchase.
The Court determined that both settlement classes satisfied the requirements of Rule 23(a) and Rule 23(b)(3). Although the Court had previously denied certification of the Privacy Class, it explained that certification standards are applied differently to litigation classes and settlement classes. The Court also recognized that the factual issues underlying its earlier denial were no longer relevant for settlement purposes.
The Court then evaluated the settlement under Federal Rule of Civil Procedure 23(e), which requires that class action settlements be fair, reasonable, and adequate.
The Court examined the adequacy of the $68 million recovery, particularly in light of the risks of continued litigation. For the Purchaser Class, Plaintiffs estimated that a successful trial could produce damages ranging from approximately $120 million to $448.8 million. Thus, the settlement represented approximately 15.15% to 56.67% of the estimated recoverable damages.
The Privacy Class faced even greater litigation risks because it would first need to successfully appeal the denial of class certification before obtaining classwide relief. Accordingly, the Court found that the different recovery amounts between the two classes were reasonable and equitable.
The Court also considered the reaction of class members. As of September 30, 2026, approximately 1.49 million timely claims had been submitted, compared to only 217 valid and timely requests for exclusion and seven objections. The Court found that the overwhelmingly positive response strongly supported approval of the settlement.
Ultimately, the Court concluded that the settlement was fair, reasonable, and adequate, particularly considering the uncertainty, expense, and delay associated with continued litigation.
Proposed Award of Attorneys’ Fees
Class counsel requested $22,666,666.70 in attorneys’ fees, representing approximately one-third of the $68 million settlement fund, along with $1,021,738.40 in litigation expenses.
The Court evaluated the request using the two methods recognized in the Ninth Circuit: (1) the percentage-of-recovery method and (2) the lodestar method.
Under the percentage method, courts generally apply a 25% benchmark in common-fund cases, which may be adjusted based on the circumstances of the litigation. Here, the Court found that the substantial recovery, complexity of the case, litigation risks, and extensive work performed by counsel supported an upward adjustment.
The Court however, concluded that a lodestar cross-check was appropriate. Class counsel reported spending approximately 20,371.6 hours litigating the case, resulting in a claimed lodestar of $22,477,337.50. But while the Court was impressed with the work performed, it wasn’t quite as impressed with some of the hourly rates. The Court analyzed the requested hourly rates, found several to be unreasonable and therefore adjusted those rates. These adjustments essentially reduced the lodestar by $217,800. Nevertheless, the Court found that the revised lodestar cross-check ratio of approximately 1.009 continued to support the requested fee award.
Thus, despite reducing several hourly rates for purposes of the lodestar calculation, the Court ultimately approved the full $22,666,666.70 attorneys’ fee request, finding that an award representing one-third of the settlement fund was fair and reasonable.
The Court also approved the requested $1,021,738.40 in litigation expenses, which included expert and consultant fees, mediation expenses, electronic discovery costs, document hosting, court reporting, and other expenses incurred throughout the litigation.
Service Awards and Objections
The four class representatives requested service awards of $10,000 each in recognition of their participation in the litigation.
However, the Court found those amounts unreasonable when compared to the relatively modest recoveries available to other class members.
Specifically, the Court observed that a $10,000 service award would exceed the highest estimated recovery for a Purchaser Class member by more than 178 times. Although the Court acknowledged the representatives’ substantial contributions, including participating in lengthy depositions and assisting counsel throughout the litigation, it concluded that $5,000 per representative was more appropriate and therefore reduced them.
The Court also addressed seven objections concerning various aspects of the settlement, including the amount of individual compensation, attorneys’ fees, claim submission requirements, and adequacy of notice. Ultimately, the Court overruled all seven objections, granted final approval of the settlement, awarded $22,666,666.70 in attorneys’ fees and $1,021,738.40 in expenses, and approved $5,000 service awards for each of the four class representatives.
One thing’s for sure: Google’s alleged listening habits came with a hefty price tag! As always, stay tuned for the latest developments!
