Corporate & Commercial
Cite as: Karen Alexander Horowitz, When is § 230 Immunity Lost?: The Transformation from Website Owner to Information Content Provider, 3 Shidler J. L. Com. & Tech. 14 (Apr. 6, 2007), at <http://www.lctjournal.washington.edu/Vol3/a014Horowitz.html>
© 2007 Karen Alexander Horowitz
Section 230 of the Communications Decency Act (CDA) offers broad protection for providers and users of interactive computer services against liability for defamation and other content-based claims when a third-party provides the information. Although providers and users of interactive computer services (ICSs) are permitted to exercise some editorial control while still avoiding legal liability, at some point, such editing may transform the provider or user of the ICS into an information content provider and deprive them of § 230 immunity. The key issue is where the threshold between permissible and impermissible editorial control lies. This Article delves into this issue by analyzing two recent decisions by the U.S. Court of Appeals for the Eleventh Circuit, Whitney Information Network v. Xcentric Ventures and Almeida v. Amazon.com, Inc., which indicate that even relatively minor editing of content could deprive the provider or user of the ICS of § 230 immunity.
Introduction
Broad Interpretation of § 230 of the CDA
Decisions Prior to the Passage of § 230
Provisions of § 230
Decisions After the Passage of § 230
Whitney Information Network v. Xcentric Ventures
Almeida v. Amazon.com, Inc.
Significance of the Eleventh Circuit’s Recent Decisions
Conclusion
Practice Pointers
<1> In Whitney Information Network, Inc. v. Xcentric Ventures, L.L.C. 2 the United States Court of Appeals for the Eleventh Circuit indicated that simply inserting a word, which may add emphasis to the online text without altering the meaning, might be enough to remove the provider or user of an interactive computer service3 (ICS) from the protections granted by Congress under the Communications Decency Act § 230 immunity.4 The case raises serious questions about which party bears the burden of proof in establishing whether the provider or user of the ICS acted as a content provider.5
<2> Since it was enacted in 1996, courts have interpreted § 230 to immunize Internet service providers (ISPs) and providers or users of ICSs from tort violations committed by users over their systems.6 To lose § 230 immunity and become liable for information posted by a third-party, providers or users of ICSs generally have to make a significant contribution to the third-party material that alters its meaning.7 The majority of circuits have stated that simply in exercising the traditional editing functions, such as deciding whether to publish, withdraw, postpone, or alter content, providers or users of an ICS did not transform from service provider into a content provider.8 While providers or users of ICSs are protected from liability by § 230, content providers are not.9 However, Whitney raises questions about the extent of editing that is permissible under § 230.10
<3> Moreover, dicta found in the recent case, Almeida v. Amazon.com, Inc.,11 also indicates that the Eleventh Circuit may be interpreting § 230 immunity to apply less broadly than the other circuits.12 In this case, the Eleventh Circuit appears to be defining the line between ISP and content provider by examining whether the provider filters or censors the information, in addition to analyzing the role of the provider or user of the ICS in editing content.13
<4> This Article will first examine some of the problems providers or users of ICSs face in making forums available for their users, and case law on ICS liability before the passage of § 230. Then the Article will turn to existing case law indicating that courts have interpreted § 230 of the CDA to confer broad immunity. This case law suggests that providers or users of ICSs may perform an array of editorial functions while maintaining their status as such and retaining their § 230 immunity. Finally, this Article will examine the recent Eleventh Circuit cases Whitney and Almeida, in which the court indicates that “traditional editorial functions” may not cover certain types of ICS activities, removing providers or users of ICSs from the shelter of § 230 liability.
<5> Prior to the passage of § 230, how the traditional concept of defamation would be applied in the new Internet era remained unclear.14 The increased prevalence of Internet chat rooms, bulletin boards, and other forums for discussion led to a far greater number of potential plaintiffs with defamation claims. The anonymity of the Internet also left these plaintiffs unable to identify the original “speaker” of that information.15 The law was unable to adapt and keep up with the Internet’s explosive growth, as evidenced by two inconsistent decisions regarding providers’ or users’ of ICSs scope of liability, Cubby, Inc. v, CompuServe, Inc.16 and Stratton Oakmont, Inc. v. Prodigy Services Co.,17 which occurred during the Internet’s earlier years.
<6> First, in 1991, the United States District Court for the Southern District of New York in Cubby held that a provider or user of an ICS could only be held liable if it “knew or had reason to know” of the allegedly defamatory statements.18 An individual who alleged that a publication available on one of CompuServe’s bulletin boards defamed him sued the corporation. The court stated that CompuServe was acting as a distributor of third-party information, similar to a library, bookstore, or newsvendor, and found that once CompuServe agreed to carry a particular publication, it had little or no control over the editorial content. CompuServe was accordingly allowed the heightened standard of liability applicable to distributors.19
<7> Second, in an unpublished 1995 decision, the New York Supreme Court in Stratton held that Prodigy could be liable for an allegedly defamatory message that an unidentified user had posted on a Prodigy bulletin board.20 While the court accepted the Cubby court’s ruling that a distributor may not be held liable unless it had notice of the allegedly defamatory content, it held that Prodigy was not the equivalent of a mere distributor. Because Prodigy had advertised that it exercised control over third-party content, the court concluded that Prodigy had “opened it[self] up to greater liability than CompuServe and other computer networks” by failing to remove the alleged defamatory material.21 Prodigy, the Stratton court held, was more akin to a traditional publisher, such as a newspaper, because it exercised editorial control over its content.
<8> Combined, these two decisions led to a rather bizarre legal state of affairs, prompting Congressional action.22 A provider or user of an ICS who did not attempt to exercise editorial control and thus contributed to the lingering presence of defamatory information would not be held liable (Cubby),23 while a provider or user of an ICS who did seek to exercise control would face potential liability (Stratton).24 In response to the uncertainty created by these two decisions, 25 combined with the desire to promote the growth of the Internet while simultaneously restricting access to objectionable material, Congress passed § 230.
<9> The Communications Decency Act became effective February 8, 1996.26 The CDA added § 230 to the Communications Act of 1934,27 and took a decidedly different approach towards libel online. Perhaps due to fears that the Stratton decision would serve as a disincentive toward exercising editorial control, Congress overruled Stratton with § 230(c), labeled “Protection for ‘Good Samaritan’ blocking and screening of offensive material.”28 Under § 230(c)(1), “[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.”29
<10> Congress also specifically sought to immunize providers and users of ICSs from liability for attempting to restrict access to objectionable material.30 Section 230(c)(2) states that no provider or user of an ICS shall be held liable for any action taken in good faith to restrict access to obscene material. The language of the statute indicates that Congress had at least two motives behind its actions. First, Congress sought to help the embryonic online industry flourish.31 Second, Congress sought to encourage providers or users of ICSs to screen the content they provide and to make available to their users the means to limit their exposure to certain types of material.32 In an attempt to ensure that Congress’ policy judgments become the law of the land, § 230(e)(3) states that “[n]o cause of action may be brought and no liability may be imposed under any State or local law that is inconsistent with this section.”33
<11> Since the immunity conferred by § 230(c)(1) is based on the defendant’s status as a provider or user of an ICS, and appears to apply even to publications made in bad faith, it is generally seen as an absolute privilege against liability for third-party speech.34 Section 230(c)(2) is also generally viewed as conferring an absolute privilege; however, because § 230(c)(2)(A) applies only to actions “voluntarily taken in good faith,” it appears to create a qualified privilege.
<12> The statutory language of § 230 of the CDA does not explicitly define the threshold where a provider or user of an ICS is transformed into an information content provider through the exercise of editorial control. It is difficult to determine what constitutes being “responsible” for the creation or development of information from the language alone.35 Accordingly, providers or users of ICSs should look to the case history to determine what constitutes permissible editing.
<13> In Zeran v. America Online, Inc., the Fourth Circuit explicitly stated that simply exercising the traditional editing functions, such as deciding whether to publish, withdraw, postpone or alter content, is not enough to transform an individual from a provider or user of an ICS to an information content provider.36 The court stated that § 230 creates federal immunity to any cause of action that would make service providers liable for information originating with a third-party user of the service.37 In reaching this holding, the Fourth Circuit emphasized the threat that tort-based lawsuit pose to freedom of speech in the Internet era and that § 230 was enacted, in part, to promote the continued growth of the Internet.38
<14> Following the Fourth Circuit in Zeran, in 2004, the U.S. District Court for the District of Columbia reached a similar decision in Ramey v. Darkside.39 The case involved a dancer who sued a website publisher after one of its advertisers used a photo of her without her permission. In its decision, the court stated that, “because defendant did no more than select and make minor alterations to the advertisement, it cannot, as a matter of law, be considered the content provider of the advertisement.”40
<15> Moving beyond the minor degree of editorial control discussed in Ramey, the courts seem to have interpreted § 230 immunity to extend even where a provider or user of an ICS has exercised a significant amount of editorial control. For example, in Ben Ezra, Weinstein & Co. v. America Online, the Tenth Circuit held that even though AOL occasionally corrected errors in stock quotations that appeared on its proprietary network, AOL did not contribute to the “development or creation of the stock quotation information.”41 Although AOL could arguably have been considered “responsible” for the information posted since it was corrected based on AOL’s notification, the Tenth Circuit held that this was not enough to transform AOL into an information content provider.42
<16> Similarly, in Schneider v. Amazon.com, Inc.,43 the Washington Court of Appeals held that Amazon was not liable for negative comments posted on Amazon’s site by third parties about plaintiff and his books. Although Amazon required that postings satisfy the company’s guidelines, and reserved the right to edit and/or remove such postings, the court held that Amazon was a provider of an ICS entitled to avail itself of the protections afforded by § 230.44 In particular, the court stated that Amazon was immunized from the claims in question because they premised Amazon’s liability on its failure to remove offending content originated by others, an editorial function for which the statute was intended to provide protection.45
<17> The courts have gone a step beyond Zeran, Ben Ezra and Schneider, and have held that providing a format for third-party content is also not enough to make a provider or user of an ICS an information content provider for the given form information. In Carafano v. Metrosplash.com, Inc.,46 the Ninth Circuit held that an Internet matchmaker was not an information content provider even though the website had provided a dating questionnaire.47 The Ninth Circuit made the distinction that since the selection of the content was left exclusively to the user, the fact that some of the content was formulated in response to the questionnaire was irrelevant.48 Similarly, eBay has created a highly structured feedback system that provides a specific format for users to enter content and that categorizes each customer response. In Gentry v. eBay,49 the California Court of Appeals held that the feedback system did not transform eBay into an information content provider because eBay did not create or developing the underlying information.50
<18> Despite holding that § 230 immunity applies broadly, the courts have recognized that once a provider or user of an ICS, through its edits, significantly alters the meaning of the third-party information, the provider or user of the ICS becomes the information content provider for that altered content.51 For example, a provider or user of an ICS is clearly not acting simply as an editor if it alters the statement “Fred is not a criminal” by removing the word “not.” However, the courts have never clarified the line between acceptable editing and the point at which a provider or user of an ICS is transformed into an information content provider. That line may fall fairly close to the original content, according to the Eleventh Circuit.
<19> In Whitney Information Network, Inc. v. Xcentric Ventures, L.L.C.,52 the Eleventh Circuit, in dicta, interpreted § 230 to apply less broadly than the other circuits and adopted a view that was more like the New York Supreme Court’s opinion in Stratton.53 Although this case cannot be said to constitute a true circuit split because there are no other cases that contradict the Whitney holding, two key elements of the Eleventh Circuit’s opinion indicate that this court may be taking a different view of § 230 immunity than the other circuits. First, the Eleventh Circuit indicated that inserting words within third-party content, which do not change the overall meaning but do add emphasis, might be enough to deprive a provider or user of an ICS of § 230 immunity.54 Second, and perhaps more significantly, the Eleventh Circuit shifted the burden of proving that it was not an information content provider back onto the provider or user of the ICS.55
<20> On its surface, Whitney is a classic § 230 case. The defendants operate a website, ripoffreport.com, where consumers can submit complaints about businesses. A user posted a negative critique of the plaintiff, who runs real estate training programs, on the website. Afterwards, the negative critique was indexed by search engines, which resulted in users accessing critical opinions of the plaintiff when they searched for the plaintiff’s name or product. Subsequently, the plaintiff sued for defamation.
<21> The defendants were acting within the scope of § 230 immunity when they allowed a negative critique to be posted on ripoffreport.com. However, the plaintiff also alleged that the defendants altered the content of the critique by inserting words such as “ripoff”, “dishonest”, and “scam.”56 The plaintiff asserts that adding such words removed ripoffreport.com from the protections of § 230.
<22> In response, the defendants did not try to argue whether or not such revisions would transform them into an information content provider; instead, they conceded that inserting those words would constitute an “active participation” on their part in generating the alleged defamatory content. As their defense, the operators of ripoffreport.com presented affidavits, which tended to show that they did not modify any of the posted content. The website operators asserted that their policy was to review the reports before posting solely to redact profanity, obscenity, and personal contact information.
<23> The Eleventh Circuit, however, was not satisfied with the defendants’ affidavits, and found that they were not adequate to shift the burden of proof back to the plaintiffs. Since the defendants now had the job of proving that they acted as providers or users of ICSs, and not content providers, the case was remanded for further proceedings.57 This holding has two key implications: first, if the plaintiff is in fact able to show that the defendants had added words such as “rip-off” and “scam” to the third-party postings, Xcentric would no longer have § 230 immunity; and second, the defendants bear the burden of rebutting the allegation that they acted as information content providers.58
<24> By implying in dicta that Xcentric would no longer have § 230 immunity if found to have inserted words into the third-party content, the Eleventh Circuit is providing insight into its view of the scope of § 230 immunity and what it considers to constitute a “traditional editorial function.” This case sits between past cases — it is more than simply providing a form or questionnaire, but less than drastically altering the meaning of the posted content. As such, the Eleventh Circuit’s decision in Whitney could be said to be a clarification of the middle ground between acceptable and unacceptable editing.
<25> Some scholars, however, argue that the Eleventh Circuit is deviating from the other circuits with its holding in Whitney.59 The crux of this argument is that Xcentric’s editing essentially amounted to simply refining the content for style, and that doing so is part of the traditional editorial functions. By implying that Xcentric’s actions would be enough to deprive them of their § 230 immunity, these scholars argue that the Eleventh Circuit is taking too narrow a view of the scope of § 230 immunity.
<26> This conclusion is supported by the manner in which the Eleventh Circuit referenced the definition of “information content provider.” The Eleventh Circuit notes that the CDA defines an “information content provider” as “any person or entity that is responsible, in whole or in part, for the creation or development of information provided through the Internet or any other interactive computer service.”60 This emphasis on partial responsibility for the content indicates that, unlike the courts which construe § 230 to provide broad immunity, that the Eleventh Circuit may be more likely to hold a provider or user of an ICS liable when it has made even a small contribution to the defamatory content.
<27> The Eleventh Circuit’s burden-shifting analysis in Whitney is also extremely significant; it illustrates that the court may be less sympathetic than other circuits toward providers and users of ICSs. By placing the burden of showing that they did not act as information content providers on Xcentric, the Eleventh Circuit potentially exposed providers and users of ICSs to greater responsibility in litigating § 230 claims. Under the Eleventh Circuit’s view of § 230 immunity, while providers and users of ICSs are certainly still not liable if they do not act as content providers, the loss of that presumption is still significant and points towards a narrowing of the scope of § 230 immunity.
<28> Combined with their interpretation of what qualifies as a “traditional editorial function” and the emphasis on being “partially responsible” for the content, this burden shifting indicates an overall unwillingness by the 11th Circuit to interpret § 230 immunity as broadly as is seen in the earlier cases. This could serve as a warning to interactive computer services that this court might be more likely to find that exercising editorial control could transform it into an information content provider.
<29> Dicta found in another recent Eleventh Circuit case, Almeida v. Amazon.com, Inc., 61 also indicates that the court may be interpreting § 230 immunity to apply less broadly than the other circuits. The Eleventh Circuit in Almeida ultimately bases its decision on grounds other than § 230 immunity, but makes several statements in a footnote, which implies that it has a different interpretation of § 230 than the other circuits.62
<30> The Eleventh Circuit references the other circuits’ interpretation of § 230 immunity by stating that, “the majority of circuits also read subsection (c)(1), though phrased as a definition, to block civil liability when interactive service providers refrain from filtering or censoring the information on their sites.”63 This statement is significant for two reasons. First, it implies that the Eleventh Circuit believes that the subsection is properly read as a definition rather than as a general conferral of immunity. Second, it implies that a provider or user of an ICS would be transformed into an information content provider if the provider or user of the ICS did not refrain from filtering or censoring the information. Other circuits have interpreted § 230(c)(2)(a) to allow providers or users of ICSs to take good faith actions to limit access to objectionable content without being exposed to civil liability.64 Thus, not only is the Eleventh Circuit displaying a general disdain for the common interpretation of § 230 as a conferral of immunity, but it is also limiting the breadth of § 230 immunity by implying that censoring or filtering information would transform a provider or user of an ICS into an information content provider.
<31> The Eleventh Circuit also notes in the same footnote that, “as a factual matter, there is no indication that Amazon had knowledge of the allegedly misappropriated image.” Although the other circuits have repeatedly asserted that actual or constructive knowledge is irrelevant to a § 230 determination,65 this dicta implies that the Eleventh Circuit may be looking to limit the application of § 230 immunity and might give some weight to whether or not the provider or user of the ICS was aware of the third-party content being posted. Providers or users of ICSs should be aware that under the Eleventh Circuit’s interpretation of § 230, they may be more likely to be exposed to liability for third-party content.
<32> While the Eleventh Circuit has indicated in dicta that it may not interpret § 230 as a broad conferral of immunity, other circuits have continued to hold that § 230 does indeed shield providers or users from tort violations committed by users over their systems. In fact, there is some evidence that the scope of § 230 immunity is still expanding.
<33> An example of the expansion of § 230 immunity can be seen by examining the courts’ treatment of two separate classes of defendants. In 1998, in Blumenthal v. Drudge, the U.S. District Court for the District of Columbia held that § 230 immunity bars claims against “interactive computer services” (such as AOL) for information provided by another content provider.66 As seen in subsequent cases, such as Ben Ezra, this became the standard interpretation of § 230. In 2006, the California Supreme Court, in Barrett v. Rosenthal, became the first court to broaden § 230’s reach and hold that § 230 defamation immunity also extended to an individual Internet “user” who is not a provider.67
<34> On the other hand, the Eleventh Circuit in Whitney and Almeida has not been the only court to question the proper interpretation of § 230 immunity. In Chicago Lawyers’ Comm. For Civil Rights Under the Law v. Craigslist, Inc., the Northern District of Illinois also questioned in dicta prior courts’ interpretations of § 230.68 In particular, the court questioned the holding of Zeran v. America Online, identifying three problems with its holding.69 First, the court stated that Zeran overstated the plain language of the statute when it held that § 230 creates a federal immunity to any cause of action that would make service providers liable for information originating with third-party users of the service.70 Second, the court found Zeran’s holding to be internally inconsistent.71 Third, the court noted that the application of the statute would be problematic, inasmuch as the policy of encouraging providers of online computer services to police for objectionable content is at odds with the immunity that would attach to providers that choose to do nothing to filter objectionable content.72 Despite the court’s concerns with the shortfalls of Zeran’s interpretation of § 230, the court ultimately held that Craigslist was entitled to immunity, and noted that plaintiffs who attempt to hold ICSs liable for content provided by others still have “a tough row to hoe.”73
<35> In the face of competing court decisions regarding the scope of § 230 immunity, the significance of the dicta found in the Eleventh Circuit’s decisions in Whitney and Almeida should not be overstated. While these cases may be an indication that the Eleventh Circuit is interpreting § 230 to apply more narrowly than other courts or even possibly indicative of an emerging trend in § 230 immunity, they might also be examples of some dissatisfaction with the current state of the law. The true significance of Whitney and Almeida has yet to be determined.
<36> Generally speaking, courts have interpreted § 230 immunity to apply broadly. As long as a provider or user of an ICS does not stray from the traditional editing functions, he does not become an information content provider and is not liable for third-party content. The courts, however, have never precisely indicated where the line is between simply editing someone else’s content and being responsible for providing some of that content. Recently, the Eleventh Circuit, in two separate decisions, has shown a disinclination to interpret § 230 as a broad conferral of immunity and has indicated that tailoring a third-party’s content to serve your purposes, while still maintaining the overall meaning of the text, may transform a provider or user of an ICS into an information content provider. This may simply be a clarification of the middle ground between permissible and impermissible editing, or it may be indicative of a desire to limit the scope of § 230 immunity. The Eleventh Circuits’ decisions in Whitney and Almeida occurred during a time in which some legal scholars have stated that § 230 should be amended to narrow the immunity it confers on certain types of websites.74 Accordingly, website owners should be on notice that the Eleventh Circuit’s decision may be part of a larger trend of removing some of the protections of § 230.