State of N.C. v. Tiktok Inc.

State of N.C. v. TikTok Inc., 2025 NCBC 47. STATE OF NORTH CAROLINA IN THE GENERAL COURT OF JUSTICE SUPERIOR COURT DIVISION WAKE COUNTY 24CV032063-910 STATE OF NORTH CAROLINA, ex rel. JEFFREY N. JACKSON, ATTORNEY GENERAL, Plaintiff, v. ORDER AND OPINION TIKTOK INC.; TIKTOK U.S. DATA SECURITY INC.; TIKTOK LLC; ON DEFENDANTS’ TIKTOK PTE. LTD.; TIKTOK, LTD.; MOTIONS TO DISMISS BYTEDANCE INC.; and BYTEDANCE LTD., Defendants. 1. Nationwide, social-media companies are defending lawsuits claiming that their apps are addictive and harmful to minors. In this case, the State of North Carolina has sued the owners and operators of TikTok, a popular app for sharing and viewing user-created videos. According to the State, the makers of TikTok designed the app to be highly addictive to minors and then undertook a deceptive publicity campaign to convince parents and children that the app is safe. On that basis, the State asserts a claim for unfair or deceptive trade practices under N.C.G.S. § 75-1.1. 2. Pending are the defendants’ motions to dismiss on several jurisdictional and merits-related grounds. For the following reasons, the Court DENIES the motions. North Carolina Department of Justice, by Joshua D. Abram, Charles G. White, Kunal Choksi, and Brian D. Rabinovitz, for Plaintiff State of North Carolina, ex rel. Jeffrey N. Jackson, Attorney General. Brooks, Pierce, McLendon, Humphrey & Leonard LLP, by Eric M. David, Gabrielle Motsinger, Jennifer K. Van Zant, and William Gregory Gaught, and O’Melveny & Myers, LLP, by Daniel M. Petrocelli, Lauren F. Kaplan, Jonathan Hacker, Martha F. Hutton, and Stephen D. Brody, for Defendants TikTok Inc., TikTok U.S. Data Security Inc., TikTok LLC, TikTok Pte. Ltd., TikTok Ltd., ByteDance Inc., and ByteDance Ltd. Conrad, Judge. I. BACKGROUND 3. The following background assumes that the allegations in the complaint are true. 4. TikTok is a software application that allows users to view, post, and interact with short-form videos. The defendants are ByteDance Ltd. and six subsidiaries bearing some variant of the name ByteDance or TikTok. Together, they own, design, market, and operate the app. For simplicity, the Court will refer to the app as TikTok and the defendant companies, collectively, as ByteDance. (See Compl. ¶¶ 12–18, 22, 25, 26, 33, ECF No. 3.) 5. It would be hard to overstate TikTok’s popularity, especially with minors. The app first launched in the United States in 2018. Just two short years later, nearly every American teen with a smartphone was using the app at least monthly. In North Carolina alone, almost a million teens aged thirteen to seventeen and hundreds of thousands of younger children used the app as of 2023. (See Compl. ¶¶ 32, 38, 41.) 6. According to the State, ByteDance targeted teens and children to fuel its growth. The main way that ByteDance makes money is through advertising, which means that revenue goes up when users spend more time on TikTok. Early on, ByteDance identified teens as a “key user segment[] to grow” because they were more engaged and likely to continue to use the app into adulthood. Indeed, minors were viewed as a dependable audience because they, unlike adults, “do not have executive function to control their screen time.” (Compl. ¶¶ 34, 49, 50, 118, 119.) 7. TikTok features an array of elements allegedly designed to exploit minors’ developmental immaturity and induce compulsive use. TikTok’s home page (coined the “For You Page”) feeds each end user videos that are algorithmically selected to maximize engagement. The algorithm, or recommendation system, performs this task by recording the user’s interactions with the app (such as sharing or skipping a video), identifying behavioral patterns, comparing the user’s behavior with others’, and ranking videos as more or less likely to be engaging based on that comparison. This individualized feed is, in the words of ByteDance employees, “addictive.” (Compl. ¶¶ 49, 57–59, 118, 119.) 8. Other design elements enhance TikTok’s addictive quality. When a user opens TikTok, a video plays automatically. The user can then cycle through videos endlessly just by swiping a finger. These features—“autoplay” and “infinite scroll”— generate an immersive, seamless experience without the occasional pause that the user might regard as a natural stopping point. Of course, scrolling isn’t all that the app has to offer. Filters allow users to touch up photos and videos in myriad ways; one filter called “Beauty Mode” makes facial features and hairstyles look more attractive. Various buttons and widgets also allow users to like and share videos, post comments, and follow specific content creators. The desire to amass likes and similar social rewards begets more frequent and protracted app usage. (See Compl. ¶¶ 70–72, 86, 87, 93, 96.) 9. In addition, the app sends push notifications to coax users to return to the app when they are away. Notifications arrive on a schedule most likely to get users’ attention, such as late in the evening. They may highlight algorithmically selected videos and sometimes promote content that is available to view only for a short period or at a specific time, playing on users’ fear of missing out to create a sense of urgency. There are also badges, which appear as a number above the app’s icon and tempt the user to return by quantifying, perhaps falsely, all that they’ve missed while not using the app. (See Compl. ¶¶ 73, 77–82.) 10. These design choices allegedly make TikTok addictive to minors in much the same way that, say, roulette is addictive to gamblers. One reason that roulette is so alluring is that it offers unpredictable, variable rewards. As the wheel spins, gamblers “anticipate[] a reward that they know could come but is tantalisingly just out of reach,” and they “experience a dopamine rush” in the process. TikTok has similar traits. Each new video in the feed is a surprise; users can scroll as long as they wish, endlessly anticipating but never knowing what they will see next. By the same token, social rewards are variable; users do not know when they will get the next notification that a viewer followed their account or liked one of their videos. For minors, this is irresistible. As one expert put it, minors “struggle ‘to ignore the prospect of a dopamine reward, even when this conflicts with other essential daily activities, such as sleeping or eating.’ ” Tracking statistics bear this out: the average teen user opens TikTok sixteen times per day, and many teens spend more than four hours on the app daily and often wallow in lengthy, late-night binges. (Compl. ¶¶ 52, 54, 55, 62, 87, 104, 105, 113.) 11. TikTok addiction is bad for minors’ mental health, the State alleges. ByteDance’s own employees have sounded the alarm to company leaders, worrying that compulsive use of the app disturbs sleep patterns, interferes with “work/school responsibilities,” “leads to deficient self-regulation,” causes or compounds “anxiety,” and impairs “analytical and problem-solving skill[s], memory formation, contextual thinking, conversational depth and empathy.” Beyond that, employees have expressed concerns about negative effects on teens’ self-image and susceptibility to eating disorders, particularly in connection with appearance-altering filters. But not everyone at ByteDance shares these concerns. One dissenter brushed them aside, asking “isn’t addiction in this sense considered a very positive metric in our field?” And ByteDance’s leadership has allegedly rejected these and other internal calls to make the app less addictive. (Compl. ¶¶ 98, 101, 108, 109, 114–17, 123–128.) 12. To the outside world, though, ByteDance touts its safety efforts. In advertisements and public statements, ByteDance maintains that minor accounts carry an automatic screen-time limit of sixty minutes, built-in nudges to remind minors to take breaks, and customizable tools that parents can use to control what their kids see and do. Plus, ByteDance tells users that they can escape so-called “rabbit holes” of entrancing, personalized content by refreshing their feed as if they had just opened a new account. In the same vein, ByteDance publicizes Community Guidelines that draw the lines between permissible and impermissible content. The Guidelines forbid and promise the removal of, among other things, “sexually suggestive” content created by minors and images of “gory, graphic human injuries.” (Compl. ¶¶ 136–38, 147, 150, 161, 175, 179, 187, 189, 195.) 13. But these are smokescreens, according to the State. In every case, the purported safety features either do not function as advertised (refreshing the feed lasts just a few videos before heading back down the “rabbit hole”) or are so easy to disable as to be useless (teens get a passcode that they can use to override the supposed limit on screen time). Indeed, ByteDance allegedly designed these features to be ineffective while giving the illusion of mitigating compulsive use. And ByteDance disregards its own Community Guidelines. Its skeletal content-moderation staff cannot review more than a small fraction of posted content for compliance, so that scads of videos and user comments that are “egregious,” “dangerous,” and “graphic” by ByteDance’s self-imposed standards never get flagged. Even when aware of prohibited content, ByteDance chooses to make that content harder to find but not to remove it as promised. (See, e.g., Compl. ¶¶ 139, 140, 156, 166, 168, 178, 182, 188, 205, 206, 211–14.) 14. The State asserts a single claim, based on TikTok’s design and marketing, for unfair or deceptive trade practices under N.C.G.S. § 75-1.1. It claims that ByteDance unfairly designed TikTok to be addictive to minors despite knowledge that compulsive use harms them. It also claims that ByteDance deceived the public by misrepresenting TikTok’s safety features and Community Guidelines while falsely assuring that the app is safe for young users. (See, e.g., Compl. ¶ 219.) 15. ByteDance has moved to dismiss the complaint, both for lack of personal jurisdiction and for failure to state a claim for relief. (See ECF Nos. 7, 35.) The motions are fully briefed, and the Court held a hearing on 24 April 2025. The motions are ripe for decision. II. PERSONAL JURISDICTION 16. North Carolina is not home to any defendant. One is based and incorporated in Singapore, two are based in China and incorporated in the Cayman Islands, and four are based in California and incorporated in either California or Delaware. (See Compl. ¶¶ 12–18.) Together, they contend that the Court lacks personal jurisdiction over them. Because both sides have offered affidavits and exhibits, the Court “must determine the weight and sufficiency of the evidence before it” in assessing whether jurisdiction exists. Toshiba Glob. Com. Sols., Inc. v. Smart & Final Stores LLC, 381 N.C. 692, 694 (2022). 17. “Personal jurisdiction refers to the Court’s ability to assert judicial power over the parties and bind them by its adjudication.” In re A.B.D., 173 N.C. App. 77, 83 (2005) (citation and quotation marks omitted). Jurisdiction must be authorized by the State’s long-arm statute, N.C.G.S. § 1-75.4, and consistent with the Due Process Clause of the Fourteenth Amendment to the United States Constitution. See Beem USA LLLP v. Grax Consulting, LLC, 373 N.C. 297, 302 (2020). In practice, this “two-step analysis” usually collapses into one because our Supreme Court has broadly construed the long-arm statute “to make available to the North Carolina courts the full jurisdictional powers permissible under federal due process.” Id. (citation and quotation marks omitted). 18. Due process requires that a defendant “have certain minimum contacts” with this State “such that the maintenance of the suit does not offend ‘traditional notions of fair play and substantial justice.’ ” Int’l Shoe Co. v. Wash., 326 U.S. 310, 316 (1945) (quoting Milliken v. Meyer, 311 U.S. 457, 463 (1940)). Although courts “have differentiated between general or all-purpose jurisdiction, and specific or case-linked jurisdiction,” only the latter is at issue here. Goodyear Dunlop Tires Operations, S.A. v. Brown, 564 U.S. 915, 919 (2011). Specific jurisdiction depends on the “relationship among the defendant, the forum, and the litigation.” Daimler AG v. Bauman, 571 U.S. 117, 133 (2014) (citation and quotation marks omitted). There must be “an affiliation between the forum and the underlying controversy, principally, activity or an occurrence that takes place in the forum State and is therefore subject to the State’s regulation.” Goodyear, 564 U.S. at 919 (cleaned up). Put more succinctly, the plaintiff’s claims “must arise out of or relate to the defendant’s contacts with the forum.” Ford Motor Co. v. Mont. Eighth Jud. Dist. Ct., 592 U.S. 351, 359 (2021) (cleaned up). 19. In addition, the “defendant’s contacts with the forum state must be such that [it] should reasonably anticipate being haled into court there.” Beem USA, 373 N.C. at 303 (cleaned up). The defendant “must take some act by which it purposefully avails itself of the privilege of conducting activities within the forum State.” Ford, 592 U.S. at 359 (cleaned up). “Random, fortuitous, or attenuated contacts” won’t do. Walden v. Fiore, 571 U.S. 277, 286 (2014) (cleaned up). Rather, the material contacts “must show that the defendant deliberately reached out beyond its home—by, for example, exploiting a market in the forum State or entering a contractual relationship centered there.” Ford, 592 U.S. at 359 (cleaned up). 20. Here, ByteDance has extensive, purposeful contacts with North Carolina. It markets the TikTok app throughout the State while cultivating ongoing relationships with a vast body of in-state users. Millions of North Carolinians, including hundreds of thousands of teens and children, have downloaded and used TikTok. In doing so, they have accepted the app’s privacy policies and terms of use, which broadly allow ByteDance to collect their data. And the reason that ByteDance collects data from North Carolina users is so that it can identify their preferences and sell advertising tailored to them. (See Am. White Decl. ¶¶ 4, 5, 22, 24, 30, ECF No. 53; Compl. ¶¶ 40– 44, 118, 119.) 21. ByteDance downplays these contacts. 1 In its words, “merely providing a generally available online platform that is accessible to a state’s residents does not subject a defendant to personal jurisdiction in that state.” (Br. Supp. 6, ECF No. 36.) 1 In its reply brief, ByteDance—for the first time—accuses the State of improper group pleading. This argument comes too late. The complaint expressly alleges that “[a]ll Defendants operate together as a common enterprise” and that “agency and/or alter-ego relationships have formed.” (Compl. ¶ 19; see also Compl. ¶¶ 20–25 (alleging “interchangeable” operations and “blurred” company boundaries).) It also alleges that “each Defendant has actively formulated, participated in, approved, directed, or otherwise controlled the acts or practices referenced throughout this complaint and has jointly advertised, marketed, developed, and distributed the TikTok social media application and platform to consumers in North Carolina.” (Compl. ¶ 26.) If ByteDance wished to rebut these allegations or argue that they are deficient, it should have done so in its opening brief. Because it did not, the argument is waived, and the allegations are deemed uncontested. See BCR 7.7 (“[T]he Court may decline to consider issues or arguments raised by the moving party for the first time in a reply brief.”). In addition, it denies “that any marketing in North Carolina was specifically targeted toward North Carolina.” (Br. Supp. 6.) 22. To be sure, courts have long held “that a defendant’s maintenance of a passive website does not support the exercise of personal jurisdiction over that defendant in a particular forum just because the website can be accessed there.” Jennings v. AC Hydraulic A/S, 383 F.3d 546, 550 (7th Cir. 2004). But TikTok isn’t a passive website. It is an interactive software application—a product that most users download on their smartphones and carry with them in their pockets and purses. Unlike a general web search, TikTok use is neither a one-and-done transaction nor a one-sided inquiry from a user. ByteDance regularly sends communications into North Carolina (through push notifications to users’ phones, for example) and retrieves data from in-state users (including personal data that ByteDance monetizes via targeted third-party advertisements). What results is a two-way exchange of information across state borders on a massive scale. 23. Moreover, to say that ByteDance merely makes TikTok generally available across the country would be wrong. Ample evidence shows that ByteDance actively markets the app in this State through both national and local channels. In addition to advertising online and on local television, ByteDance promotes the app to local businesses for use in their own marketing. Plus, it sponsors grants awarded to local schools by the National Congress of Parents and Teachers. Although ByteDance denies having ultimate control over the grant program, it communicated its preferred recipients and named North Carolina as a “key priority.” (Am. White Decl. ¶¶ 6–15, 22, 24, 29–31; see also Exs. C–M, R, S, X–Z to Am. White Decl., ECF Nos. 57, 51.4– .14, .19, .20, .26, .27; Kersul Decl. ¶¶ 4, 8, ECF No. 36.2.) 24. In short, ByteDance advertises widely in North Carolina, makes TikTok available here, and fosters ongoing relationships with its app’s users. All this adds up to the active, purposeful exploitation of a market in North Carolina. See Ford, 592 U.S. at 365 (observing that purposeful availment was met when Ford Motor Company had advertised “[b]y every means imaginable” in the forum, made cars “available for sale” there, and “foster[ed] ongoing connections to its cars’ owners”); Ind. v. TikTok Inc., 245 N.E.3d 681, 690 (Ind. Ct. App. 2024) (concluding with “little trouble” that Indiana’s courts could exercise jurisdiction based on similar facts, including that TikTok “knowingly and repeatedly transmit[s] data to and from each of those millions of Indiana end-users each and every hour of each and every day”). 25. Without question, ByteDance’s North Carolina-based conduct relates to the State’s claim that TikTok is addictive and harmful to minors and that ByteDance has misrepresented its efforts to make the app safe for their consumption. Indeed, ByteDance has “systematically served a market in” North Carolina “for the very” product that allegedly addicts and harms the State’s minor population. Ford, 592 U.S. at 365. “[W]hen a corporation has continuously and deliberately exploited a State’s market, it must reasonably anticipate being haled into that State’s courts to defend actions based on products causing injury there.” Id. at 364 (cleaned up). 26. That holds true even though TikTok has a national audience and even though ByteDance has exploited markets in many other States in a similar fashion. The fact that ByteDance has extensive contacts elsewhere does not diminish the extent or quality of its contacts with North Carolina. And the fact that its business is everywhere does not mean that it can be sued nowhere other than its home forum. See id. at 355 (noting that Ford’s “business is everywhere”). A contrary decision “would have the perverse effect of allowing a corporation to direct its activities toward all 50 states yet to escape specific personal jurisdiction in each of those states for claims arising from or relating to their relevant contacts in the forum state that injure that state’s residents.” Briskin v. Shopify, Inc., 135 F.4th 739, 758 (9th Cir. 2025) (en banc). 27. Accordingly, the Court denies the motion to dismiss for lack of personal jurisdiction. III. RULE 12(b)(6) 28. Satisfied that jurisdiction exists, the Court turns to ByteDance’s three remaining grounds for dismissal. First, ByteDance contends that federal law immunizes it from liability. Second, it contends that the First Amendment bars the State’s claim. Third, it contends that the State’s allegations, even if true, do not state a viable claim for relief. For these disputes, the Court must treat all well-pleaded allegations as true and view the facts and permissible inferences in the light most favorable to the State as the nonmoving party. See, e.g., Sykes v. Health Network Sols., Inc., 372 N.C. 326, 332 (2019). 29. Immunity. Under federal law, “[n]o provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.” 47 U.S.C. § 230(c)(1). In so many words, this language immunizes a provider of an internet platform from liability for any legal claim that treats it as a publisher or speaker of third-party content. See, e.g., A.B. v. Salesforce, Inc., 123 F.4th 788, 792–93 (5th Cir. 2024). 30. There’s no dispute that ByteDance provides an interactive computer service. ByteDance maintains that it is immune because the State’s theories of liability treat it as a publisher of third-party content, just as section 230 forbids. The unfairness theory, it contends, targets the way that TikTok selects, organizes, and presents content to users, and the deception theory faults it for doing too little to catch and delete harmful content. These are all core publishing functions protected by section 230, according to ByteDance. 31. The State frames things differently. It contends that the unfairness theory concerns product design, not publishing, because it is rooted in allegations that ByteDance devised technological features to make TikTok addictive to minors. And the deception theory, it contends, concerns ByteDance’s own speech, not others’, including false statements about TikTok’s safety features and Community Guidelines. In the State’s view, section 230 does not shield ByteDance from liability for its unfair design choices or its deceptive statements. 32. In recent years, federal courts have solidified section 230’s outer boundaries, stressing that it “provides internet platforms with limited legal protections.” Henderson v. Source for Pub. Data, L.P., 53 F.4th 110, 119 (4th Cir. 2022). The statute “bar[s] certain claims, in specific circumstances, against particular types of parties” but does not go so far as to “insulate a company from liability for all conduct that happens to be transmitted through the internet.” Id. at 129. Put bluntly, “this immunity is not limitless,” Calise v. Meta Platforms, Inc., 103 F.4th 732, 739 (9th Cir. 2024), and “was not meant to create a lawless no-man’s-land on the Internet,” Fair Hous. Council v. Roommates.com, LLC, 521 F.3d 1157, 1164 (9th Cir. 2008) (en banc). 33. Section 230 uses the word “publisher” in its common-law sense. “At common law, a publisher was someone who intentionally or negligently disseminated information to third parties.” Henderson, 53 F.4th at 121. “Traditional publisher liability held that if a publisher took upon itself the task of moderating or editing the content that appeared within its pages, it became responsible for anything tortious written there.” Est. of Bride v. YOLO Techs., Inc., 112 F.4th 1168, 1175 (9th Cir. 2024). In that sense, publishers were “treated differently” than mere “distributors, such as bookstores.” Calise, 103 F.4th at 739 (cleaned up). 34. Thus, when section 230 says not to treat an internet platform “as the publisher or speaker of” others’ content, it means not to burden the platform with traditional publisher liability. The statute’s reach ends there. It does not relieve internet publishers “from all potential liability” or provide “an all purpose get-out-of-jail-free card for businesses that publish user content on the internet.” Doe v. Internet Brands, Inc., 824 F.3d 846, 853 (9th Cir. 2016). And no court has indulged the “notion that a claim holds a defendant liable as a publisher anytime there is a ‘but-for’ causal relationship between the act of publication and liability.” Henderson, 53 F.4th at 122. Rather, viewed against the backdrop of the common law, “a claim only treats the defendant ‘as the publisher or speaker of any information’ under § 230(c)(1) if it (1) bases the defendant’s liability on the disseminating of information to third parties and (2) imposes liability based on the information’s improper content.” Id. at 123; see also Hill v. StubHub, Inc., 219 N.C. App. 227, 236 (2012) (“The liability must be based on the defendant having acted as a publisher or speaker.” (cleaned up)). 35. As in most areas of the law, substance matters more than form. A plaintiff may not dodge the statute through artful pleading any more than a defendant may parlay its publishing activity into blanket immunity. The inquiry is “whether the duty that the plaintiff alleges the defendant violated derives from the defendant’s status or conduct as a publisher or speaker.” Calise, 103 F.4th at 740 (cleaned up). So, for example, section 230 applies when “the duty would necessarily require an internet company to monitor, alter, or remove third-party content”—the meat and potatoes of traditional publisher liability. Salesforce, 123 F.4th at 795 (cleaned up). But it does not apply when “the duty springs from another source—for example, a contract,” Calise, 103 F.4th at 740, or “a promise or representation” made to a user, Bride, 112 F.4th at 1178. 36. Neither of the State’s theories seeks to hold ByteDance liable for monitoring, altering, or removing user content, or for failing to do those things. The thrust of the unfairness theory is that ByteDance purposely designed TikTok to be addictive to minors. If what the complaint says is true, TikTok is packed with features—autoplay, endless scrolling, social rewards, and more—that exploit minors’ developmental immaturity and neurological susceptibility to intermittent, variable rewards. And TikTok addiction allegedly disrupts healthy sleep habits and social interactions, causing insidious psychological harms to teens and children. This theory has more in common with products liability than publisher liability, resting as it does on an alleged duty not to design and offer a product that endangers a vulnerable population. See Lemmon v. Snap, Inc., 995 F.3d 1085, 1092 (9th Cir. 2021) (“The duty underlying such a [products-liability] claim differs markedly from the duties of publishers.”). 37. It is no answer to say, as ByteDance does, that addicted minors spend their time on TikTok viewing third-party content. ByteDance’s business is, after all, to host and display user videos. Nearly everything it does is connected in some way to its users’ content. But it and other social-media platforms “continue to face the prospect of liability, even for their ‘neutral tools,’ so long as plaintiffs’ claims do not blame them for the content that third parties generate with those tools.” Id. at 1094. The State’s unfairness theory neither blames ByteDance for its users’ content nor aims to hold it accountable in its capacity as a publisher of that content. The theory instead seeks to hold ByteDance liable “for its own injurious conduct” in “creating and employing tools to addict young users.” Commonwealth v. Meta Platforms, Inc., 2024 Mass. Super. LEXIS 161, at *16 (Mass. Super. Ct. Oct. 17, 2024) (denying immunity for similar anti-addiction claim against social-media company). 38. A case comparison may help. In Lemmon, grieving parents blamed the maker of Snapchat for the deaths of their sons in a high-speed vehicle crash. The boys died while using Snapchat’s “Speed Filter” to record and superimpose their driving speed onto a photo or video. Lemmon, 995 F.3d at 1088. They and many other users apparently believed that Snapchat would reward them with a digital achievement for taking a snap while driving over 100 miles per hour. See id. at 1089. The Ninth Circuit held that section 230 did not bar the parents’ negligent-design claim, reasoning that liability did “not rest on third-party content.” Id. at 1093. Rather, the claim treated the defendant “as a products manufacturer” and faulted it “solely for Snapchat’s architecture, contending that the app’s Speed Filter and reward system worked together to encourage users to drive at dangerous speeds.” Id. at 1092–93. Section 230 offered no immunity against “being sued for the predictable consequences of designing Snapchat in such a way that it allegedly encourages dangerous behavior.” Id. at 1094 (cleaned up). 39. By contrast, Bride applied section 230 to bar a product-liability claim against the maker of an anonymous messaging app. The plaintiffs had allegedly faced severe bullying and harassment by anonymous users. See Bride, 112 F.4th at 1173. They brought two sets of claims: one set based on the app maker’s false promise that it would “unmask and ban abusive users”; and a second set for product liability on the theory that the app was “inherently dangerous because of its anonymous nature.” Id. at 1178, 1180. The Ninth Circuit allowed the misrepresentation claims to proceed. Id. at 1179. But it viewed the product-liability claims as an attempt to hold the app maker “responsible for users’ speech”—that is, “faulting [the defendant] for not moderating content in some way, whether through deletion, change, or suppression” of bullying and harassing messages. Id. at 1180. Section 230 therefore immunized the defendant against the purported product-liability claims. 40. As in Lemmon, the State’s unfairness theory treats ByteDance as a product designer, not a publisher, and faults it for offering a combination of features and social rewards that foster compulsive use by minors. Unlike Bride, ByteDance’s liability does not turn on user content or its failure to remove or suppress that content. This sort of anti-addiction claim therefore does not implicate section 230, as many state courts have held. See, e.g., Dist. of Columbia v. Meta Platforms, Inc., 2024 D.C. Super. LEXIS 27, at *32 (D.C. Super. Ct. Sept. 9, 2024) (holding that section 230 does not bar claim based on “addictive design features employed by Meta—and not any particular third-party content”); State v. Meta Platforms, Inc., 2024 Vt. Super. LEXIS 146, at *15–16 (Vt. Super. Ct. July 28, 2024) (“Whether they are watching porn or puppies, the claim is that [young users] are harmed by the time spent, not by what they are seeing. The State’s claims do not turn on content, and thus are not barred by Section 230.”); In re Soc. Media Cases, 2023 Cal. Super. LEXIS 76992, at *94 (Cal. Super. Ct. Oct. 13, 2023) (“Section 230 does not bar a claim based on features of a social media site that have an adverse effect on users apart from the content of material published on the site.”). 2 2 Not all courts agree.See State ex rel. Rosenblum v. TikTok Inc., 2025 Ore. Cir. LEXIS 5135, at *36–37 (Ore. Cir. Ct. June 13, 2025) (dismissing claim under section 230 and characterizing TikTok’s tools as “content amplification features meant to publish content in a manner that increases user engagement”); In re Soc. Media Adolescent/Pers. Inj. Prods. Liab. Litig., 753 F. Supp. 3d 849, 881 (N.D. Cal. 2024) (concluding that claims based on app features stem from third-party content and are therefore barred under section 230). Some of these decisions have been criticized for employing a form of but-for reasoning. See Meta Platforms, 2024 D.C. Super. LEXIS 27, at *30. 41. The deception theory is, if anything, more straightforward. As alleged, ByteDance portrayed screen-time limits, parental tools, and other features as serious curbs on compulsive use, knowing all the while that the features did not work as advertised. At the same time, ByteDance reneged on promises made in its Community Guidelines to police user videos and comments and to remove explicit content. The legal duty at issue arises from ByteDance’s own promises and representations, not from its “failure to take certain moderation actions.” Bride, 112 F.4th at 1178. 42. ByteDance insists that this is an attempt to hold it liable for failing to identify and remove violent, sexual, and racist content. It is not. “While yes, online content is involved in these facts, and content moderation is one possible solution for [ByteDance] to fulfill its promise, the underlying duty being invoked . . . is the promise itself.” Id. at 1179 (allowing misrepresentation claims to proceed). Section 230 gives internet platforms wide latitude to moderate content. But it does not shield them from liability for breaching their promises or misrepresenting their content-moderation activities. See Meta Platforms, 2024 Mass. Super. LEXIS 161, at *11–12 (“[C]laims based on a publisher’s representations about its publishing content are not immunized under Section 230.”). 43. Accordingly, the Court denies the motion to dismiss on section 230 grounds. 44. Free Speech. The First Amendment to the federal Constitution guarantees the freedom of speech. One protected “aspect of speech” is “the editorial function”— that is, “exercising editorial discretion in the selection and presentation of content,” including third-party content. Moody v. NetChoice, LLC, 603 U.S. 707, 731 (2024) (cleaned up). Still, “the First Amendment does not prevent restrictions directed at commerce or conduct from imposing incidental burdens on speech.” Sorrell v. IMS Health Inc., 564 U.S. 552, 567 (2011) (contrasting “restrictions on protected expression” with “restrictions on economic activity or, more generally, on nonexpressive conduct”). 45. ByteDance contends that the State’s claim violates this constitutional guarantee by aiming to muzzle its editorial discretion, on the one hand, and compel speech, on the other. The unfairness theory, it says, attacks constitutionally protected editorial features that select, organize, and display videos, while the deception theory is a disguised attempt to regulate the adoption and enforcement of content-moderation standards. Going further, ByteDance worries that a judgment against it would include injunctive relief compelling it to warn consumers of TikTok’s risks to minors. 46. The State maintains that its unfairness theory concerns conduct, not speech, and that TikTok’s addictive features are not expressive. Although it acknowledges that its deception theory does concern speech, the State contends that the false and misleading commercial speech at issue is not protected by the First Amendment. The State also offers assurances that, if successful, it will seek only to enjoin ByteDance from making false statements without compelling any speech. 47. Few areas of the law are more cutting-edge than this one. It was just last year in Moody that the United States Supreme Court confirmed that “some [social-media] platforms, in at least some functions, are indeed engaged in expression” protected by the First Amendment. 603 U.S. at 716. Specifically, a platform’s content-moderation policy reflects its editorial judgments “about whether—and, if so, how—to convey posts having a certain content or viewpoint.” Id. at 738; see also id. at 717 (questioning state laws that “limit[ed] the platforms’ capacity to engage in content moderation—to filter, prioritize, and label the varied messages, videos, and other content their users wish to post”). These “editorial judgments influencing the content” of social-media feeds are “protected expressive activity.” Id. at 744. 48. “But what if,” as Justice Barrett asked in her concurrence, “a platform’s algorithm just presents automatically to each user whatever the algorithm thinks the user will like—e.g., content similar to posts with which the user previously engaged?” Id. at 746 (Barrett, J., concurring). Is that protected expressive activity as well? The Moody majority left that question unanswered: “We therefore do not deal here with feeds whose algorithms respond solely to how users act online—giving them the content they appear to want, without any regard to independent content standards.” Id. at 736 n.5. 49. This case invites Justice Barrett’s question once again. The State bases its unfairness theory on features that induce compulsive use, including TikTok’s algorithm. As alleged, the algorithm does not “understand” or “care about” content, nor does it “promote or suppress particular political agendas, views, or content.” (Compl. ¶ 60 & Fig. 2.) Rather, the algorithm presents videos based solely on an analysis of “the user’s pattern of engagement.” (Compl. ¶ 59.) Put another way, “[t]he recommendation engine is content-neutral, meaning that it recommends content based on behavioral and certain device signals, not on the semantic nature of the content itself.” (Compl. ¶ 60 & Fig. 2.) 50. Taking these allegations as true, it’s hard to discern any expressive activity. The algorithm does not convey a message by its programmer; it simply bows to user preferences and propensities. At a minimum, the complaint supports an inference that a reasonable person would understand TikTok’s video feed to reflect a given user’s content choices as opposed to ByteDance’s own creative expression or editorial judgment. See, e.g., U.S. Telecom Ass’n v. FCC, 825 F.3d 674, 743 (D.C. Cir. 2016) (“As a result, when a subscriber uses her broadband service to access internet content of her own choosing, she does not understand the accessed content to reflect her broadband provider’s editorial judgment or viewpoint.”); see also NetChoice v. Bonta, 761 F. Supp. 3d 1202, 1222 (N.D. Cal. 2024) (“[I]t would be hard to say that the algorithm reflects any message from its creator because it would recommend and amplify both favored and disfavored messages alike so long as doing so prompts users to spend longer on social media.”). 51. So too for the other disputed design features, such as autoplay, infinite scrolling, and social rewards. ByteDance says little about them, offering no basis to conclude that they deserve First Amendment protection independent of the underlying algorithm. And on the face of the complaint, none of the features has an obviously expressive quality. See Holloman ex rel. Holloman v. Harland, 370 F.3d 1252, 1270 (11th Cir. 2004) (“Thus, in determining whether conduct is expressive, we ask whether the reasonable person would interpret it as some sort of message, not whether an observer would necessarily infer a specific message.” (emphasis omitted)). 52. Of course, the complaint gives only a glimpse—and likely a contested glimpse at that—into the inner workings of TikTok’s algorithm and ancillary features. A more developed record might reveal an expressive message that changes the calculus. But for now, the Court concludes that the First Amendment does not mandate dismissal of the State’s unfairness theory. See Meta Platforms, 2024 Mass. Super. LEXIS 161, at *22–23 (deeming claims about features that “allegedly induce addiction in young users . . . to be principally based on conduct and product design, not expressive content”); see also Meta Platforms, 2024 D.C. Super. LEXIS 27, at *43 (same); Meta Platforms, 2024 Vt. Super. LEXIS 146, at *17–18 (same); Utah Div. of Consumer Prot. v. TikTok Inc., No. 230907634, at 14 (Utah Dist. Ct. Nov. 12, 2024) (same). 53. Likewise, the Court concludes that the First Amendment does not bar the State’s deception theory. ByteDance misconstrues the theory as an attempt to regulate its content-moderation practices to suppress overly engaging content. That is not what is alleged. Rather, the State alleges that ByteDance deceived the public by making false and misleading statements about the functionality of its safety features and its adherence to its Community Guidelines. “Untruthful speech, commercial or otherwise, has never been protected for its own sake.” Va. State Bd. of Pharm. v. Va. Citizens Consumer Council, 425 U.S. 748, 771 (1976); see also Hest Techs., Inc. v. State ex rel. Perdue, 366 N.C. 289, 297 (2012) (observing that fraudulent speech “receive[s] no First Amendment protection”). 54. To round things out, ByteDance need not fear the prospect of compelled speech. There is, as yet, no judgment against it. If the State prevails and seeks an injunction compelling ByteDance to speak, only then will the issue become ripe. 55. Accordingly, the Court denies the motion to dismiss on First Amendment grounds. 56. Failure to State a Claim. The General Assembly has “declared unlawful” all “unfair or deceptive acts or practices in or affecting commerce,” N.C.G.S. § 75-1.1, and empowered the State’s attorney general to prosecute violators “whenever in his opinion the interests of the public require it,” id. § 75-15. Modeled after section 5 of the Federal Trade Commission Act, section 75-1.1 regards a practice as “unfair when it offends established public policy as well as when the practice is immoral, unethical, oppressive, unscrupulous, or substantially injurious to consumers.” Marshall v. Miller, 302 N.C. 539, 548 (1981). “A practice is deceptive if it has the capacity or tendency to deceive; proof of actual deception is not required.” Winston Realty Co. v. G.H.G., Inc., 314 N.C. 90, 98 (1985) (cleaned up). 57. ByteDance argues that its alleged actions, even if true, are neither unfair nor deceptive. It is not unfair, ByteDance contends, to offer an otherwise lawful product 3 that is extraordinarily engaging, especially when consumers remain free to 3 The United States Supreme Court dealt with a challenge to legislation affecting TikTok’s legality earlier this year. See TikTok Inc. v. Garland, 145 S. Ct. 57, 62 (2025) (per curiam) (“As of January 19, the Protecting Americans from Foreign Adversary Controlled Applications Act will make it unlawful for companies in the United States to provide services choose other forms of entertainment. As for deception, ByteDance argues that the State has not alleged supporting facts with particularity and that any supposed misrepresentations weren’t false or were mere puffery or statements of opinion. 58. The State accuses ByteDance of ignoring allegations that it designed TikTok to exploit the neurological vulnerabilities of minors, fully aware of the harms caused by compulsive use. This exploitation of minors, the State contends, is unfair and offensive to public policy. In addition, the State continues, ByteDance engaged in a public campaign to placate concerned parents with deceptive statements about its content-moderation activities and the app’s safety features, thus doubly violating section 75-1.1. 59. This case is by no means the first attempt to hold a social-media company accountable for unfair practices under consumer-protection laws. Many States have pursued similar claims against social-media titans (including ByteDance) under their own consumer-protection statutes, also modeled after the FTC Act with slight variations. In those cases, one court after another has concluded that designing an app to induce addictive, compulsive use by minors comfortably qualifies as an unfair practice. The federal court overseeing multidistrict litigation against Meta (the company behind Facebook) held that exploiting minors’ developmental susceptibility to addictive features met every state-specific standard of unfairness as well as the “purportedly stricter” federal standard. In re Soc. Media Adolescent Addiction/Pers. to distribute, maintain, or update the social media platform TikTok, unless U.S. operation of the platform is severed from Chinese control.”). Inj. Prods. Liab. Litig., 753 F. Supp. 3d at 899. 4 These decisions involving analogous laws and comparable allegations are highly persuasive. See Marshall, 302 N.C. at 542 (noting that “decisions interpreting the FTC Act may be used as guidance in determining the scope and meaning of” section 75-1.1). 60. ByteDance’s chief argument—that consumers can make a free and informed decision to use or not use TikTok—does not reckon with the complaint’s allegations. The consumers in this scenario are minors, not mature adults. As alleged, ByteDance designed TikTok to exploit the unique vulnerabilities that accompany youthful immaturity, inducing addictive, compulsive use and depriving minors of a free choice. Worse yet, ByteDance did so allegedly knowing that addiction causes social and psychological harms to minors. Taken as true, these allegations outline practices that are “immoral, unethical, oppressive, unscrupulous, or substantially injurious” under section 75-1.1. Marshall, 302 N.C. at 548; see also, e.g., In re Soc. Media Adolescent Addiction/Personal Inj. Prods. Liab. Litig., 753 F. Supp. 3d at 897 (questioning whether “a minor still retains that ‘free and informed choice’ ” when “the minor has developed—by the platform’s own design—a compulsive habit to click through and continue use”). 61. Turning to deception, ByteDance’s particularity argument is unpersuasive. “The degree of particularity required . . . varies from case to case.” Haigh v. Superior 4 The federal court rejected challenges to claims brought under at least six state consumer-protection statutes. See id. at 902 (upholding claims by California, Colorado, Indiana, Kansas, New Jersey, and Oregon). Individual actions in state courts have followed suit. See, e.g., Meta Platforms, 2024 Mass. Super. LEXIS 161, at *25–26 (finding allegations that social-media company designed an app to have “features that it knows encourage addictive use by teenagers” unfair under consumer-protection statute). Ins. Mgmt. Grp., 2017 NCBC LEXIS 100, at *27 (N.C. Super. Ct. Oct. 24, 2017) (citation and quotation marks omitted). Here, the State alleges that ByteDance conducted a yearslong campaign on several fronts to convince the public that TikTok is safe for minors. The complaint identifies many false and misleading statements in press releases, advertisements, congressional testimony, and TikTok’s Community Guidelines over that period—statements that ByteDance conspicuously skips over in its briefing. (See, e.g., Compl. ¶¶ 138, 153, 187, 203.) These allegations are sufficiently particular. 62. Moreover, the allegedly deceptive statements are not vague, indefinite, or puffery. Three examples will suffice. First, ByteDance advertises that TikTok automatically limits teens’ screen time when, in fact, the app has no daily limit. (See Compl. ¶¶ 137, 139.) Second, ByteDance tells users that they may refresh their feed to get a “fresh start on TikTok,” but this feature worked only temporarily for most users and not at all for others. (Compl. ¶¶ 152, 156, 158.) Third, ByteDance represents that it removes content that violates its Community Guidelines but in practice chooses not to enforce the guidelines at all or to make illicit content harder to find without removing it. (See, e.g., Compl. ¶¶ 187, 188, 197.) In each case, the alleged statement is a definite representation about how features work or how ByteDance moderates content. See, e.g., In re Soc. Media Adolescent Addiction/Pers. Inj. Prods. Liab. Litig., 753 F. Supp. 3d at 891 (denying motion to dismiss deception claim based on “yearslong course of conduct replete with numerous public statements regarding user safety”). 63. Elsewhere, the complaint details ByteDance’s statements that it “prioritizes the safety” of minors and offers “practical tools” to manage their screen time. (Compl. ¶¶ 163, 164.) ByteDance urges that these and similar statements are too subjective to support a deception claim. But the complaint goes on to allege that ByteDance didn’t mean what it said. While touting TikTok’s safety tools in public, ByteDance was allegedly telling its employees in private to design those features so that they would be hard to find, hard to use, and broadly ineffective. (See, e.g., Compl. ¶¶ 144, 156, 158, 159, 183.) Product designers understood that “[o]ur goal is not to reduce the time spent” by minors on the app. (Compl. ¶ 168.) Even a statement of opinion may support a deception claim when the speaker “holds an opinion contrary to the opinion” being expressed. Leftwich v. Gaines, 134 N.C. App. 502, 509 (1999); see also, e.g., Meta Platforms, 2024 Vt. Super. LEXIS 146, at *21 (holding that representation that “Instagram is safe” was actionable when taken with specific allegations that Meta designed Instagram to induce compulsive use and denied research showing its addictiveness). 64. Finally, bound as it is to stay within the four corners of the complaint, ByteDance has not shown as a matter of law that the disputed statements were true. ByteDance argues, among other things, that it paired its promise to remove content violating the Community Guidelines with a caution that it “cannot guarantee that all content shared on TikTok complies.” (Br. Supp. 32–33.) That is hardly a silver bullet. The State’s theory is not that ByteDance muffed an impossible task (to catch and remove all offensive conduct). It is that ByteDance promised to remove illicit content and made no good-faith attempt to follow through. Whether ByteDance’s myriad statements were true or false is a matter for discovery. 65. Accordingly, taking the allegations as true and resolving all inferences in favor of the State, the complaint adequately states a claim under section 75-1.1. The Court therefore denies the motion to dismiss for failure to state a claim. IV. CONCLUSION 66. If the State’s allegations are true, ByteDance has intentionally addicted millions of children to a product that is known to disrupt cognitive development, to cause anxiety, depression, and sleep deprivation, and (in the worst cases) to exacerbate the risk of self-harm. Federal law does not immunize this conduct, the First Amendment does not bless it, and North Carolina’s laws and courts are not powerless to address it. 67. For all these reasons, the Court DENIES ByteDance’s motions to dismiss. SO ORDERED, this the 19th day of August, 2025. /s/ Adam M. Conrad Adam M. Conrad Special Superior Court Judge for Complex Business Cases