Content moderation is one of the biggest challenges for any interactive computer service provider that publishes information created by others. It is difficult to balance users' desires to express themselves and their desires to avoid unpleasant or even hateful speech. The providers' ultimate tool to stop users from ruining the computer service for others is to ban those users from publishing content on the service. But what if those users keep coming back? And what if they organize others to violate the platform's terms of service on another platform?
On September 9, 2021, Twitch Interactive, Inc., tried to solve this problem by suing two pseudonymous users, CruzzControl and CreatineOverdose,1 for carrying out numerous "hate raids" on its platform.2 A "hate raid" is conducted when a malicious actor uses Twitch's "raid" feature to flood another streamer with hateful messages, often with the help of automated bot accounts.3
In its complaint, Twitch admitted that it has been unsuccessful in investigating and banning those two users, who would create new accounts after being banned.4 Twitch believes the users have built software code to perform automated hate raids and have updated the code to avoid Twitch's efforts to ban their automated bot accounts.5
Among other things, Twitch alleged two causes of action that are unusual for interactive computer service providers to bring as plaintiffs against individual users: breach of contract for and fraudulent inducement claims under California common law.6
Breach of Contract
Twitch's breach of contract claim is an interesting twist on how online platforms defend themselves against bots and trolls. Usually, platforms allege breach of contract defensively when users challenge punishments after their content is moderated. For example, in Murphy v. Twitter, Inc., a freelance journalist, Murphy, sued Twitter for breach of contract (among other claims) after the platform locked her account for violating its hateful conduct rules.7 Twitter defeated Murphy's claim because Twitter's terms of service expressly state that it can suspend or terminate accounts for any reason at its discretion, without liability.8
In other cases, platforms sue users for egregious misuses of the platform that improperly seek to monetize the platform. For instance, in Twitter, Inc. v. Skootle Corp., Twitter sued users for violating Twitter's terms of service by using automated accounts to spam other users.9 After Twitter's claim survived a motion to dismiss,10 the defendants entered into a consent order, promising not to violate the Terms of Service.11 In Craigslist, Inc. v. 3Taps Inc., Craigslist alleged that 3Taps and others harvested and reproduced large quantities of Craigslist's content without authorization.12 After Craigslist's breach of contract claim survived a motion to dismiss,13 Craigslist also obtained final judgments and permanent injunctions.14
In contrast with those allegations of breach of contract, Twitch's claim is more unusual. After all, in most situations, it would be efficient and effective simply to ban users. Indeed, it is almost certain that Twitch's costs to bring the lawsuit will exceed any actual recovery it might obtain from the defendants. It seems likely, then, that this lawsuit is intended to send a message. It will be interesting to see if Twitch can obtain a deterrent effect from its prosecution of this suit.
Fraudulent Inducement
Another relatively novel claim is Twitch's claim against the users for fraudulent inducement. To prove fraudulent inducement under California law, a plaintiff must prove the following five elements: "(a) misrepresentation (false representation, concealment, or nondisclosure); (b) knowledge of falsity (or 'scienter'); (c) intent to defraud, i.e., to induce reliance; (d) justifiable reliance; and (e) resulting damage."15
As with breach of contract, when these claims are brought, they are usually brought by users against platforms (and often fail to survive a motion to dismiss). For instance, in Kidstar v. Facebook, Inc., a user alleged that Facebook fraudulently induced him to host information on Facebook, which Facebook subsequently deleted.16 On March 23, 2021, the court dismissed the plaintiff's claim for failure to point to any specific misrepresentation.17 The plaintiff's attempt to amend also failed.18 In Facebook, Inc. v. BrandTotal Ltd., BrandTotal filed a counterclaim alleging that Facebook violated California's Unfair Competition Law (UCL) through two fraudulent representations.19 The court held that the alleged correction to the first misrepresentation was in the same document and, in any event, BrandTotal was required to allege a statement likely to deceive the public under the UCL.20
One of the most challenging parts of a fraudulent inducement claim is showing that a party never intended to perform at the time it entered into the contract. The mere failure to perform does not support the inference that a party acted fraudulently: "If plaintiff adduces no further evidence of fraudulent intent than proof of nonperformance of an oral promise, he will never reach a jury."21
Here, Twitch alleges that CruzzControl and CreatineOverdose knew their hate raids violated the Terms of Service and deliberately created new accounts to continue to violate the Terms of Service. If true, this likely meets the scienter requirement for fraudulent inducement. Again, it is unlikely that the recovery from these individual defendants would be substantial. But if Twitch's lawsuit reduces the number of users evading bans by registering for new accounts, other platforms may bring similar claims to achieve deterrent effects.
Of course, lawsuits alone are not enough for interactive computer service providers, who should continually update their policies to address new threats. On September 29, 2021, Twitch announced policy changes to address the threat of "hate raids."22
Waen Vejjajiva and Brian Levy contributed to the preparation of this advisory.
[1] See Wilson Sonsini Goodrich & Rosati, Client Alert, “Pokémon Company Caught Pseudonymous Online Users but It Is Not Always So Easy,” (July 20,2021), https://www.wsgr.com/en/insights/pokemon-company-caught-pseudonymous-online-users-but-it-is-not-always-so-easy.html (describing the difficulties in obtaining the identities of pseudonymous users).
[2] See Complaint ¶ 7-8, Twitch Interactive, Inc. v. CruzzControl et al., No. 3:21-cv-07006 (N.D. Cal. Sept. 9, 2021).
[3] Id. ¶ 3.
[4] Id. ¶ 51.
[5] Id.
[6] Id. ¶¶ 58-87.
[7] Murphy v. Twitter, Inc., 60 Cal. App. 5th 12, 22 (2021).
[8] Id. at 35.
[9] Twitter, Inc. v. Skootle Corp., No. C 12-1721 SI, 2012 U.S. Dist. LEXIS 87029, at *6 (N.D. Cal. June 22, 2012).
[10] Id. at *3.
[11] See Stipulated Consent Order, Twitter, Inc. v. Skootle Corp., No. 3:12-cv-01721-JST, (N.D. Cal. June 6, 2013), ECF No. 93.
[12] Craigslist Inc. v. 3Taps Inc., 942 F. Supp. 2d 962, 966–67 (N.D. Cal 2013).
[13] Id. at 977.
[14] Orders, Craigslist Inc. v. 3Taps Inc., 3:12-cv-03816-CRB, (N.D. Cal June 30, 2015), ECF Nos. 271–272.
[15] Lazar v. Superior Court, 909 P.2d 981, 984–85 (Cal. 1996).
[16] Kidstar v. Facebook, Inc., No. 20-cv-05408-SK, 2021 U.S. Dist. LEXIS 54904, at *2 (N.D. Cal. Mar. 23, 2021).
[17] Id.
[18] See Order, Kidstar v. Facebook, Inc., No. 20-cv-05408-SK (N.D. Cal. June 9, 2021), ECF No. 77.
[19] Facebook, Inc. v. BrandTotal Ltd., 499 F. Supp. 3d 720, 730-31 (N.D. Cal. Nov. 2, 2020).
[20] Id. at 743.
[21] Tenzer v. Superscope, Inc., 702 P.2d 212, 219 (Cal. 1985).
[22] Ana Diaz, Twitch Announces New Features to Curb Hate Raids, Polygon (Sept. 29, 2021), https://www.polygon.com/22700564/twitch-hate-raids-safety-feature-streaming.
