TikTok Inc. asked a federal court in California on November 13, 2025 to declare unconstitutional, as applied to the company, a state law that bars personalized feeds for users under 18 unless a parent gives verified consent. The complaint, filed against California Attorney General Rob Bonta, argues that the For You feed is TikTok's own editorial speech. Nearly a year later, that argument has lost its first round. A judge in San Jose refused to block the law in August 2026, TikTok's appeal now sits before the Ninth Circuit, and the company has since gone to court against a second, stricter California law.

In Short

California passed a law saying apps like TikTok cannot show under-18 users a feed tailored to what they like unless a parent verifies and agrees, with a default limit of one hour a day for that feed, and TikTok sued to stop it. The fight is over whether picking videos for each person counts as free speech, much like a newspaper picking stories, and it affects the creators, brands and advertisers who reach young audiences through TikTok's recommendations. For now a judge has let the law stand, so TikTok has to follow it in California while it appeals, and from January 2027 apps will also have to check whether their California users are adults.

What TikTok asked the court to do

The complaint challenges three parts of California's Protecting Our Kids from Social Media Addiction Act, passed in 2024 as Senate Bill 976 and codified at sections 27000 to 27007 of the California Health and Safety Code. TikTok groups them as the "personalized-feed provisions." Section 27001(a) makes it unlawful for a covered operator to provide an "addictive feed" to a minor unless it has obtained verifiable parental consent. Section 27002(b)(2) requires a parental tool, switched on by default, that limits a minor's access to such feeds to one hour per day unless the verified parent changes it. Section 27002(b)(4) requires a tool that lets a verified parent make the default feed one that is not personalized at all.

According to the complaint, all three are "plainly unconstitutional as applied to TikTok." The company asked for a declaration to that effect, a preliminary and a permanent injunction barring the Attorney General from enforcing them against TikTok, and attorneys' fees. The Act gives the Attorney General exclusive authority to enforce it, which is why Bonta is the only defendant, named in his official capacity.

The filing went further than the three provisions. In a second count, TikTok asked the court to confirm that four other parts of the law, already blocked in earlier litigation brought by the trade group NetChoice, cannot be enforced against it either. Those are the restrictions on sending notifications to minors during certain hours, the companion notification tool, the default that hides like counts, and the requirement that operators publicly disclose how many minors use their service each year. The complaint notes that neither earlier decision limited its injunction to NetChoice and its members.

TikTok Inc., a California corporation based in Culver City, filed through O'Melveny & Myers, with Daniel M. Petrocelli as signing counsel. The complaint carries case number 4:25-cv-09789. The court's later orders list it as 5:25-cv-09789-EJD, before Judge Edward J. Davila in the San Jose division. Meta Platforms and Google, with YouTube, filed their own challenges the same day, which PPC Land covered on November 14, 2025. That report focused on the Google and Meta filings. TikTok's complaint sets out in unusual detail how the company says its feeds are assembled.

The scale TikTok put on the record

The numbers come early. The complaint says the platform has "more than 170 million monthly users in the United States and 1 billion users worldwide." In 2025 alone, according to the filing, people in the United States uploaded more than 7 billion videos, which were viewed more than 17 trillion times in the United States and abroad. "No single person could possibly sift through all of that content, much less identify content that they would enjoy," the complaint states. "So TikTok curates that content for them."

Moderation figures follow. Between April and June 2025, more than 189 million videos were removed for violating TikTok's Community Guidelines, according to the complaint, which cites the company's quarterly enforcement report. Over the same period, more than 99% of removed content was taken down before anyone reported it, and 90% of removed videos were never viewed by a single user.

On the reach side, the filing says the majority of video views on the platform happen through the For You feed. It describes organic discovery as the reason creators join: content can reach people beyond a creator's followers "without any paid promotion," and "it is not unusual for videos created by everyday people to receive thousands, or even millions, of views."

How TikTok says the For You feed is built

The core of the complaint is a description of three processes TikTok says it uses to compile each person's feed. All three, according to the filing, "involve human decisions about what content to publish, recommend, and promote," with technology used to apply those decisions at scale. The complaint puts it in terms of old media: "the use of technology to effectively implement editorial judgments no more strips those judgments of their constitutional protections than would the use of a printing press instead of a pen."

Content moderation

The first layer is content moderation against the Community Guidelines, which the complaint dates as effective September 13, 2025. The guidelines rest on eight principles, including preventing harm and supporting expression, and TikTok says it drafts them with outside experts, including its U.S. Content Advisory Council and what it calls "the industry's first Youth Council."

Every upload passes through an automated moderation system before it appears, and during that review the video is visible only to the person who posted it. Clear violations, such as graphic violence, are removed automatically. Ambiguous cases go to human moderators. Content that breaks no rule but "may be unsuitable for a broader audience" can be made ineligible for the For You feed or pushed down in search, and content tagged as potentially inappropriate for under-18 users becomes invisible to anyone who registers as under 18.

Some examples are specific. Videos of adults smoking cigarettes do not violate the guidelines, according to the complaint, but cannot be recommended in the For You feed. Distressing videos on matters of public interest sit behind "opt-in viewing screens" and are also excluded from recommendation.

The recommendation system

The second layer is what the complaint calls TikTok's "proprietary recommendation system," a sorting and ranking mechanism that engineers would describe as a recommender system. In the United States, the filing says, it is deployed by TikTok U.S. Data Security Inc., a special-purpose subsidiary that controls access to protected US user data.

The process, as described, runs in two steps. A pool of candidate videos is drawn from content already cleared for recommendation, then the system scores and ranks those candidates for the individual. The signals include engagement data (playtime, likes, shares, accounts followed, comments and content created), account and device information (language preference and country setting) and video information (captions, sounds and hashtags). One sentence is likely to interest anyone who buys media on the platform: "TikTok may adjust how the system considers different factors depending on, among other things, how a factor will further TikTok's mission and business goals."

The complaint is careful to say that user preferences do not run the feed alone. "While individual preferences are considered, they are not the sole factor that determines the composition of any individual's For You feed." People can steer what they see with a "Not interested" button, a feed refresh, filter words and subject preferences, and TikTok advises that scrolling quickly past a video signals disinterest. The app also explains recommendations with reasons such as "This post is popular in your country" and "This video is longer, and you seem to like longer videos."

Promotion and filtering

The third layer is video promotion and filtering. Employees hand-select video categories they want to promote, such as Super Bowl highlights or footage from a Taylor Swift concert, and each US promotion request is manually reviewed by a US-based reviewer, according to the filing.

Filtering rules work the other way. TikTok says it prioritizes content from the user's own country, avoids duplicates, enforces appropriate video length and disperses runs of videos on the same subject. Some categories are spread out deliberately because they are fine in small doses but "may not be healthy to review excessively." The complaint names extreme exercise and mental health videos as examples.

Eight feeds, not one

The For You feed is not the only target. The complaint lists eight other surfaces it says the law would reach: the Following feed, the Friends feed, Stories Skylight, the LIVE feed, the Search feed and search recommendations inside comments, the STEM feed, the Explore feed and the Comment feed, which ranks comments on a video. Each, according to TikTok, combines a person's information and interactions with the company's editorial decisions.

What SB 976 actually covers

Much of the complaint is an argument about definitions. The Act applies to an "addictive internet-based service or application," meaning a service that offers an "addictive feed" as a significant part of what it provides. An addictive feed, under section 27000.5(a), is any service or portion of one in which media "generated or shared by users" are recommended, selected or prioritized for display based on information provided by the user or associated with the user or the user's device. Seven exemptions apply, including feeds that use information not persistently tied to the user, content the user "expressly and unambiguously requested," and direct private messages.

Two kinds of service are excluded altogether: those where interactions between users are limited to commercial transactions or consumer reviews, and those whose feed exists mainly for cloud storage. A "minor" is anyone under 18 located in California.

"Despite the Act's provocative labeling, the definition of '[a]ddictive feed' does not depend in any way on addiction," the complaint argues. A footnote adds that there is no consensus among academics or professionals on what constitutes addiction in a social media context, and that the Act does not define it.

The complaint also quotes the other side. It cites the Attorney General as having said the law was needed because "the default ... algorithmic feed" offered to social media users "too often sends them down rabbit holes of harmful content." It notes that the bill faced opposition from ACLU California Action, Chamber of Progress, LGBT Tech, the Trevor Project and the Woodhull Freedom Foundation, some of which warned that parental-consent rules could cut LGBTQ+ youth off from supportive communities.

The January 1, 2027 trigger

A footnote in the complaint flags a date that matters to everyone, not only teenagers. Until January 1, 2027, the parental-consent requirement applies only to users a platform knows are minors. After that date, an operator must "reasonably determine" that a user is not a minor before serving a personalized feed without consent. "Doing so will require verifying users' age, which will affect all users, not just minors," the complaint states. Courts in the NetChoice litigation held that challenges to this age assurance requirement were not yet ripe, and TikTok's complaint does not ask the court to rule on it directly.

TikTok's own teen defaults

A large section of the complaint describes tools TikTok says already give parents the oversight the state wants. That list is central to its claim that the law is not the least restrictive option available.

The most striking overlap is time. According to the complaint, accounts registered as under 18 have Daily Screen Time switched on by default at 60 minutes; once the limit is reached, a teen must enter a passcode to keep going. SB 976's default is also one hour, though it applies only to personalized feeds and can be changed only by a verified parent.

Other defaults described in the filing include:

  • Family Pairing, which links a parent's account to a teen's to set screen time limits, mute notifications, filter keywords, turn on Restricted Mode, control search, set the account to private and restrict or switch off direct messages.
  • Mandatory notification mute windows that users cannot change: 9 p.m. to 8 a.m. for users registered as 13 to 15, and 10 p.m. to 8 a.m. for 16 and 17 year olds. A parent's less restrictive schedule is overridden by the default.
  • Sleep Hours, which interrupts the For You feed with a full-screen prompt and calming music for under-16 users after 10 p.m., followed by a second prompt that is harder to dismiss.
  • Private accounts by default for everyone registered as under 18, with the setting locked for those under 16.
  • Search interventions that direct users searching terms such as "suicide" to support resources.

The complaint also points to controls outside TikTok: carrier tools from Verizon and Xfinity, browser and device controls from Microsoft, Google, Apple and Samsung, and third-party apps such as Aura and Bark. It names them as examples of less restrictive alternatives. It does not claim any of them can switch off feed personalization inside TikTok.

The constitutional argument

TikTok's legal theory rests on the Supreme Court's 2024 decision in Moody v. NetChoice, which said that compiling and organizing third-party speech is expressive activity in its own right. Quoting that decision, the complaint argues that of the billions of posts and videos "(plus advertisements)" that could end up in a customized feed, the platform ensures that "only the tiniest fraction do." It then extends the logic: "The State can no more dictate how platforms make such editorial decisions than it can tell a newspaper which articles to present or how to present them."

From there the complaint argues that strict scrutiny applies, because the provisions single out one kind of speaker (services that personalize feeds of user-generated media) and one kind of content (feeds shaped by user preferences). It says the law is overinclusive, covering political, educational and news videos and treating a 13-year-old and a 17-year-old identically. It also says the law is underinclusive. Netflix and the New York Times can recommend content based on past use without restriction, according to the filing, and "an e-commerce website could serve teens an unlimited feed of products likely to catch their eye, but TikTok's provision of the same feed (absent parental consent) would violate the Act."

The parental-consent mechanism draws the sharpest language. The complaint calls it a prior restraint and argues that "the Act deputizes parents, guardians, and other caregivers to act as agents of the government applying whatever standards the parents decide to apply." It also takes aim at the logic of allowing the feeds with consent at all: "Teens cannot, for example, buy alcohol with parental consent." On the law's addiction framing, the filing is blunter still. "The Act's labeling is merely a sleight-of-hand that exposes that its means are not narrowly tailored to its ends."

There is a concession too. "TikTok recognizes and shares the State's interest in protecting minors, but that general interest does not support the particular prohibitions here," the complaint states.

The filing illustrates the stakes with a single user. "A seventeen-year-old who goes to TikTok to learn new things, encounter new communities, and find new voices will be deprived of the primary way TikTok facilitates that discovery, unless and until he or she gets parental consent to do so. That, simply put, is unconstitutional."

What the court did with it

The state answered with an omnibus opposition to all three companies' motions, filed on February 20, 2026. It argued that no human being reviews or selects the compilation in any individual's personalized feed, that the algorithms have no understanding of the content they compile, and that SB 976 is content-neutral because it regulates digital mechanisms rather than topics, ideas or viewpoints.

Judge Davila sided with the state at the preliminary stage. On August 5, 2026, he denied the preliminary injunction motions of TikTok, Meta and Google with YouTube in a joint order. The order found that the companies had not shown, at that stage of the case, that their personalized feeds are expressive. It described a feed driven by predicted interest as reflecting users' own perceived interests back to them, rather than conveying a message from the platform. The effect is that sections 27001(a), 27002(b)(2) and 27002(b)(4) can be enforced while the case continues.

The companies filed notices of appeal on August 11, 2026, and asked Davila for an injunction pending appeal the same day. He refused that too on August 21, 2026, after what the order describes as more than five hours of oral argument and several rounds of supplemental briefing on the original motions. The order said the companies had not shown they would suffer irreparable harm by complying while the Ninth Circuit considered their appeals, described the burden as closer to ordinary compliance costs, and noted that the state adopted the provisions to protect children's mental and physical health.

TikTok's appeal was docketed at the Ninth Circuit as No. 26-5121, with TikTok USDS Joint Venture LLC listed as a second appellant alongside TikTok Inc. The court's preliminary injunction schedule set the opening brief for September 8, 2026 and the state's answering brief for October 6, 2026. Outside groups have taken sides. The Foundation for Individual Rights and Expression filed an amicus brief urging reversal, while the Electronic Privacy Information Center and others filed one on October 20, 2026 asking the court to affirm, arguing that behavioral profiling is a functional rather than expressive part of building a feed. The schedule notice sets no date for oral argument.

Why this matters for marketers

Strip away the constitutional framing and the dispute is about who controls distribution to a large, commercially valuable audience. TikTok's own filing makes the commercial case. It says personalization, not follower counts, is how unknown creators find audiences, and it warns that a chronological feed would hand the advantage to whoever posts most. "Individual creators' posts and content will be drowned out by brands and influencers employing teams of people to post throughout the day," the complaint states. "And posters can game a chronological feed to spread more low-quality content."

What would a teenager's feed look like without personalization? TikTok's answer is that both obvious alternatives are worse for small voices: a popularity-ranked feed favors majority views, and a chronological one favors whoever posts most. If a non-personalized default spreads among California teenagers, that logic points to shrinking organic reach for small creators and a larger share for high-frequency brand accounts, at least by TikTok's account. Influencer programs built on discovery among under-18 users in the state would be the most exposed. The complaint gives no figure for how many of TikTok's US users are minors located in California.

The Europe comparison is the closest thing to evidence the filing offers. "In Europe, where a non-personalized feed option is available, only a tiny fraction of people have actually selected it," according to the complaint, which gives no figure. That option exists because the Digital Services Act requires the largest platforms to offer at least one recommender option not based on profiling. Brussels has gone further since. On February 6, 2026, the European Commission preliminarily found TikTok in breach of the Digital Services Act over addictive design, naming infinite scroll, autoplay, push notifications and its highly personalized recommender system, with possible fines of up to 6% of global annual turnover.

California has not stopped at SB 976 either. On September 10, 2026, Governor Gavin Newsom signed Assembly Bill 1709, which bars covered platforms from providing addictive features to users under 16, with civil penalties of up to $50,000 per affected minor for knowing violations. It takes effect on January 1, 2027 and, according to the complaint Google filed against it, applies even when a parent consents, unlike SB 976. Meta Platforms and Google with YouTube each sued Bonta on November 4, 2026 to block it, and TikTok filed its own challenge, according to TechCrunch, which did not give a filing date or case number for TikTok's suit. A TikTok spokesperson told the publication that AB 1709 is unconstitutional and compared personalized curation to the way bookstores, TV networks and streaming services select content. So the SB 976 appeal is no longer TikTok's only fight with the state over feeds. Meta's moves on teen accounts, including the restrictions it extended across Facebook and Messenger in April 2025, show platforms adjusting defaults on their own terms while fighting state mandates in court.

The date to watch is January 1, 2027. From then, the consent requirement no longer depends on what a platform already knows about a user's age. Any personalized feed served to a Californian without parental consent will require the operator to have reasonably determined that the user is an adult. That reaches into account creation, age signals and data handling for adults as well as teenagers, and it sits alongside long-standing federal rules such as COPPA, which already require verifiable parental consent before collecting personal data from children under 13. Whether the Ninth Circuit rules before that date is the open question. TikTok filed its complaint nearly a year ago to settle the matter; three months after a judge declined to intervene, it remains unsettled.

Timeline

Summary

Who: TikTok Inc., represented by O'Melveny & Myers, against California Attorney General Rob Bonta, with Judge Edward J. Davila of the Northern District of California and the Ninth Circuit Court of Appeals. Meta Platforms and Google with YouTube brought parallel cases.

What: A complaint seeking to block three provisions of California's Protecting Our Kids from Social Media Addiction Act (SB 976) that require verifiable parental consent before a minor gets a personalized feed, a default one-hour daily limit on such feeds, and a parental option to make the default feed non-personalized. The complaint describes how TikTok builds its For You feed and argues that curation is protected speech. The court denied preliminary relief and an injunction pending appeal, and the provisions are enforceable while TikTok appeals.

When: The complaint was filed on November 13, 2025. The preliminary injunction was denied on August 5, 2026, the appeal was filed on August 11, 2026, and the injunction pending appeal was denied on August 21, 2026. The law's age-determination requirement applies from January 1, 2027, the same day AB 1709, the newer under-16 law TikTok is also challenging, takes effect.

Where: The US District Court for the Northern District of California, San Jose division, and the US Court of Appeals for the Ninth Circuit. The law applies to users under 18 located in California.

Why: TikTok argues that personalized feeds are editorial speech and that the law singles out social platforms while leaving Netflix, news sites and e-commerce feeds untouched. California argues that algorithmic feeds are not expressive and that the law protects children's health. For marketers, the outcome decides how creators and brands reach under-18 users in California, and from 2027 how platforms establish the age of every user there.