DRAFT November 27, 2013 Mr. Kent Walker Senior Vice President and General Counsel Google Inc. Via E-mail: xxxxxxx@google.com xxxxx.xxxxxxxx@wilmerhale.com Dear Mr. Walker: Dear ____: Pursuant to our telephone conversation on Wednesday, November 20, 2013, I am sending you this letter, which I previously sent you in draft form on November 18, 2013. This letter also responds to the June 26, 2013 letter sent by Ms. Jamie Gorelick of the law firm WilmerHale on behalf of Google ("Google's Letter"). I respectfully request that you answer it in writing before our December 2, 2013, National Association of Attorneys General meeting. Based on our lengthy discussion, it is my understanding that you will neither come nor send someone with any authority to meet with some of the concerned Attorneys General during our meeting in New Orleans next week to address any of the problems that we have raised with Google. If I am wrong in my assessment of our conversation, please advise me in writing. I will advise my colleagues at the meeting next week of our conversation and I do not want to mischaracterize it. It is evident from Google's this letter , and my discussion with you, that although Google claims to be interested in cooperating with state Attorneys General, it is unwilling to take basic actions to make the Internet safe from unlawful and predatory conduct, and it has refused to modify its own behavior that facilitates and profits from unlawful conduct. In my ten years as Attorney General, I have dealt with a lot of large corporate wrongdoers. I must say that yours is the first I have encountered to have no corporate conscience for the safety of its customers, the viability of its fellow corporations or the negative economic impact on the nation which has allowed your company to flourish. I. Overview As you are aware, overwhelming evidence shows that Google facilitates and profits from numerous illegal online activities ranging from piracy to illegal drug sales and human trafficking. Yet Google has repeatedly refused to take reasonable but important steps that would reduce the ability of criminals to profit from their crimes. Google’s inaction is not merely a failure to do the right thing. Rather, it raises serious questions as to whether Google is engaged in unlawful conduct itself. Nowhere was this made more apparent than in Google’s own admissions to the United States Department of Justice when it was compelled to forfeit $500 million in profits due to its facilitation of the sale of unlawful drugs. By its own admission, Google was aware of and actively aided and abetted the unlawful sales of pharmaceuticals through its search engine and its advertising service. Notwithstanding Google’s claims that it merely provides a passive or neutral conduit for the speech of others, or the legal doctrines discussed in Google’s Letter, Google nonetheless was liable for its own conduct, its own knowledge, and its own refusal to act in the face of indisputably unlawful activity. There is every reason to believe further investigation will reveal that Google’s illegal conduct reaches far beyond the illegal pharmaceutical sales that it has already admitted to facilitating. Publicly available information, described in greater detail below, already illustrates that Google profits from a host other types of criminal conduct as well. This includes but iswell including, but not limited to, further illegal pharmaceutical sales, counterfeiting, copyright infringement, and sex trafficking. To combat these facts, Google takes two contradictory positions. Google touts its technology, especially in search – its ability to take information (including personal information about its users that raise privacy concerns not discussed in detail here) and use that information to provide “better” search results customized for individual users. With that technology and the army of Google employees who program and operate it, Google can respond to customers’ needs. It also can, as Google’s letter explains, take actions – when Google wants – against unlawful or offensive conduct by, among other things, deleting search results, removing advertising, and changing its Autocomplete feature. At the same time, Google would have us believe that it is a passive search engine incapable of combating the unlawful conduct it facilitates. That is the core of both the legal and the factual argument in Google’s Letter. Google’s admissions, however,But Google’s admissions to the United States in its $500 million forfeiture, the admissions in its letter, and Google’s own public proclamations belie such claims. Google can and does take action against unlawful or offensive conduct – when Google determines it is in its business interests to do so. GoogleBut it has chosen not to cease facilitating unlawful conduct when doing so would decrease its profits. Moreover, Google appears to gather the information it receives from separate Google products and services for purposes of enhancing its other products and services and more effectively marketing to consumers across all Google platforms. Google should apply that same philosophy to implement policies across all platforms to reduce unlawful content. For example, if Google has knowledge that a site is problematic from a Google AdSense perspective and takes action based on that, Google should share that knowledge and take the same type of action with respect to other Google products and services that may be used by or in connection with that site. As is described in greater detail below, Google can take action and does, when it so chooses. But with respect to a host of types of unlawful conduct of concern to state Attorneys General, Google simply refuses, relying on its own claimed (and false) passivity, or the First Amendment, or the technical challenges that it concedes it has overcome. To maintain its status as a legitimate business and avoid further liability, Google must finally take the actions it can to cease promoting and profiting from unlawful conduct. And it must be called to account, after a full investigation and fair hearing, for actions that are its and its alone. II. Google Facilitates And Profits From Unlawful Activities. Google does not seriously dispute that rampant unlawful conduct is assisted and made easier by Google’s search results and other conduct. Moreover, state Attorneys General, pharmaceutical companies, motion picture studios, and others have all notified Google of specific websites engaged in obviously unlawful conduct. In general, Google takes little or no action to curb the unlawful conduct, ostensibly based on the arguments made in its letter A. Search Google is unquestionably the dominant search engine on the Internet. Those who engage in unlawful activity rely heavily on Google, with Google search results being a primary, if not the primary, way that they obtain traffic to their websites. If a website selling or providing unlawful products ceased to appear at the top of Google’s search results, chosen by Google’s algorithm and its employees, it would dramatically reduce the ability of that website to violate the law and harm consumers. There is no serious dispute that Google has the capability to take such actions. As is detailed in Google’s Letter, Google removes content from its search results in a variety of circumstances: Nazi-related content is removed from search results in Germany; allegedly defamatory content is removed in the United Kingdom; insults to religion are removed in India. Google removes child pornography from its search results. It also blocks sites with spam and malware that can be damaging to users. Letter at 2. It takes these actions without waiting for a court order or court adjudication that particular content is unlawful. This is the right thing to do. Google’s successes in screening child pornography, malware, and illegal content in foreign countries demonstrate that it can curb unlawful conduct, when it so desires. Yet while Google is willing to tailor its search results to comply with the law in foreign countries, it has been unwilling to delist and demote sites that violate a variety of domestic laws. In the United States, websites with illegal content not only appear in Google’s search results, but are regularly among the top-listed search results. For example, despite the publicity surrounding Google’s facilitation of illegal and counterfeit pharmaceutical sales, when a user searches for “buy oxycodone,” the top search result on Google is a site titled “Order Oxycodone Online No Prescription.”[1] Google’s search results also facilitate piracy, forgery of identification documents, sales of counterfeit goods, cigarette sales to minors, and human trafficking. This is both troubling and inexcusable. In light of Google’s successes in screening other types of criminal content, Google cannot claim that it lacks the ability to respond to requests for assistance in fighting crime.[2] Given the obviously unlawful activities on many of these sites, once Google is notified that a site is engaged in unlawful conduct by a state Attorney General, federal law enforcement or the owner of intellectual or other property, it cannot credibly claim that it lacks knowledge. Instead, it has decided to pick and choose what unlawful conduct to combat – based on its own profit motive. Such an approach must change. One would expect that Google would be eager to act as a responsible corporate citizen and to cooperate with law enforcement to ensure that its search results do not facilitate criminal conduct. Google’s Letter claims that its “commitment to a safer internet is manifest.” Letter at 3. But this claim rings hollow in light of Google’s inaction toward the proliferation of illegal content in its search results. It appears that Google has made a calculated business decision that it will be most profitable to continue to promote websites with unlawful content and to profit from the advertising that accompanies those searches. Google should take the following actions to deal with rogue sites and discontinue the promotion of unlawful content: • Further promote authorized sites. Google should take into account information from authoritative sources on which sites have been authorized to provide content, and promote those sites in rankings for searches for that content. • Provide an icon or other indication with search results to authorized sites. Google should show an icon or other visible mark next to search results that are to known authorized sites for searches for content available on those authorized sites. • De-index rogue sites. Google should not index sites that are “rogue sites,” that is, sites substantially dedicated to intellectual property infringement. Google should de-index a site that is established to be a rogue site by referrals from trusted rights holders, or by third party services that provide meaningful criteria for assessing the level of IP infringement on websites. • Proactively refuse to index repeat infringements of content on a site. Google should revise its policies on indexing new pages on a site to content for which Google has received multiple notices of infringement on that site. • Further deprioritize rogue sites. Google should make more changes to its algorithm to push rogue sites dramatically lower in the results and to ensure that new infringing sites do not take their place. • Provide a “red light” or educational warning about rogue sites. Google should warn users before it permits them to link from Google to rogue sites. B. Autocomplete Google’s Autocomplete function, which offers real-time search term suggestions to users, only exacerbates the problems created in Google search by steering users to search for websites that engage in and promote illegal activities. Although Google’s Letter claims that Autocomplete is “analogous to automatic spell checkers” used in email programs, Letter at 3, this misrepresents the feature. Spell checking corrects objective errors that the user may have made when typing. In contrast, Autocomplete affirmatively suggests search terms to users. In essence, Autocomplete directs users to use specific terms. And Autocomplete is utterly and completely in Google’s control. While Google trumpets Autocomplete as a way to make searching faster, it also makes it far easier for users to search for and find websites that facilitate illegal activity. As recently as a few months ago, Autocomplete suggested searches for sites selling prescription drugs without prescriptions.[3] But Autocomplete goes much further, suggesting searches like “buy stolen credit card numbers,” “how to make a fake id,” and “buy bath salts drug.”[4] Autocomplete provides a road map for users to find illegal content online. Google attempts to escape responsibility for its Autocomplete suggestions by representing, “Autocomplete entries simply reflect what an algorithm predicts is the likely search query . . . .” Letter at 12. But Google ignores that it, not a third-party, authors the algorithm producing these suggestions. And althoughAlthough Autocomplete may draw from the massive data Google gathers through users’ search queries, none of the suggestions are produced by anything but Google’s algorithm. No entity other than Google could be the creator or the speaker of this content. OnceAnd, once again, it is clear that Google can control Autocomplete to combat unlawful activity, but, in most circumstances, chooses not to. Today, Google actively polices Autocomplete’s output and regularly censors that output to avoid, for example, avoid suggestions that would lead to vulgar and obscene websites. It will refuse to suggest key terms associated with child pornography. All of this is commendable. Parents should not have to worry that their children will stumble upon inappropriate content based on suggestions offered Google. But Google’s active tailoring of its Autocomplete suggestions only proves that Google, not a third party, creates and controls that content and thus promotes the searches and websites that engage in unlawful conduct. C. YouTube YouTube’s role in unlawful conduct is well-established. For example, a 2013 report by the Digital Citizens Alliance detailed how YouTube has become the how-to site for criminal behavior.[5] Videos uploaded to YouTube serve as commercials for businesses peddling prescription drugs from rogue pharmacies, forged identification documents, and counterfeit goods, as well as instructionaland as instruction videos for finding prostitutes, pirated copyrighted material, and illegal drugs.[6] But YouTube is not a mere platform for these videos. Many Rather, many of these videos are monetized to allow Google and the producer of the content to profit from the videos. When a user uploads a video to YouTube, he or she has the option to check a box to “monetize my video,” and to select the advertising formats that can accompany the video.[7] When a YouTube video is monetized, advertisements accompany the video. Google and the video’s producer share the revenues from these advertisements. When videos with illegal content are monetized, Google not only profits directly from that illegality; illegality. Worse, it allows the criminals themselves to profit. This financial partnership between Google and those who post criminal content is deeply troubling. It is also entirely avoidable. Google claims to conduct a “standard review process” before videos on YouTube are allowed to be accompanied by advertisements.[8] Such a review process certainly provides the opportunity to screen for videos that promote criminal activities. Nevertheless, these videos proliferate. What purpose does this screening process serve, if it fails to ensure that the monetized content complies with the law and YouTube’s terms of service? Google also has the ability to find these illegal videos. Google’s Letter asserts that YouTube has “created and implemented automated solutions to attempt to both remove spam videos placed by potential rogue pharmacies and to disable ads from running against videos containing metadata suggesting they might contain objectionable pharma-related content.” Letter at 5. The letter also candidly admits that after Google “learned” from news reports that videos promoting rogue pharmacies and counterfeit drugs were rampant on YouTube, it “immediately removed” thousands of videos that it found to be in violation of its guidelines. Id. This only demonstrates that Google can quickly and effectively remove videos promoting criminal activities when it chooses to do so. There is no reason why Google cannot take similar steps to address the proliferation of other criminal content on YouTube. Google should ensure that YouTube doesn’t become the go-to source for download links to popular, copyrighted content. To that end, it should remove such content from the videos and/or de-prioritize videos that contain links to known rogue sites for copyrighted content. D. Advertising Through its AdWords program, Google provides advertising and allows business – including illegitimate businesses – to promote their products and gain an advantage over other (legitimate) competitors. Even Google concedes that its legal obligations regarding the content of advertisements are heightened because they are entering into a business relationship, in many cases, with (in many cases) an obviously unlawful website. See Letter at 3. As in each of the other areas discussed above, Google concedes that it can take action to cease advertising on behalf of unlawful websites. Google’s Letter detailed various actions Google has taken to limit rogue pharmacies from advertising through its AdWords program. It is notable, however, that Google has only taken these steps under the compulsion of a Non-Prosecution Agreement with the Department of Justice.[9] With respect to a host of other types of unlawful sites, including counterfeit goods, piracy, and illegal drug sites, Google continues to do nothing. Just as the owner of a billboard would not post an advertisement informing passersby of where to buy stolen cars, Google’s AdWords program should never offer advertisements for criminal enterprises. Google must stop providing criminals with the means to target their customers and victims. III. Google Cannot Escape Liability For Its Facilitation of Unlawful Activity. The legal arguments of Google’s Letter fail for the same basic reason that Google’s factual arguments are unpersuasive – Google is not a mere passive company that just happens upon unlawful conduct, and it is not being investigated or pursued for the conduct of others. It is Google’s own conduct that renders it liable, its liable and it is Google’s conduct which must change. Firstly First, Google’s Letter argues that Google can never be liable for its support of unlawful enterprises. It claims that it could not be convicted of aiding and abetting a crime, no matter how blatant the criminal enterprise of its business partners may be. This is simply wrong. Google’s conduct has greatly exceed that of a legitimate business “provid[ing] lawful services . . . even when it is foreseeable that some small portion of users may abuse those services to promote illegal ventures.” Letter at 9. It is not merely “foreseeable” that some of Google’s advertising and YouTube partners are promoting illegal activities. To the contrary, the illegal ventures that Google facilitates are open and transparent with their conduct. Even the case law cited by Google proves why Google’s actions go beyond those of a legitimate business engaging in arms-length transactions. In Direct Sales Co. v. United States, 319 U.S. 703 (1943), the Supreme Court made clear that the type of transaction matters a great deal when determining whether a business can be liable for the criminal activities of its customer. The Court explained that some articles of commerce “from their very nature[] . . . giv[e] the seller notice the buyer will use them unlawfully.” Id. at 710. Such is the case here. When YouTube agrees to monetize a video entitled “Fake Passport USA step by step,”[10] YouTube is clearly on notice of the illegal content of the video. YouTube’s knowledge of this illegality, combined with its decision to monetize the video and share profits with its producer, proves its intent to “further, promote and cooperate” in the illegal conduct. See id. at 711. Similarly, when Google promotes, through its search results, websites obviously selling unlawful drugs or streaming pirated videos, Google cannot escape liability. Once Google it is aware of that conduct, or, in some cases, enters in a contractual relationship to promote the illegal content of the poster or advertiser, it assists criminal actors in advertising their criminal services. We recognize that the question of Google’s knowledge is key. But Google’s own admissions in the $500 million forfeiture demonstrate that Google was well aware that it was promoting unlawful conduct, took no action to stop it and, indeed, took steps to affirmatively assist the unlawful conduct – all the while earning a healthy profit. This conduct – classic aiding-and-abetting – is, we believe, likely to be replicated with respect to other forms of illegal conduct like human trafficking and product piracy. These are areas of great concern to state Attorneys General and not only relate to violations of law but also to public health and safety. Secondly Second, Google asserts that it effectively has blanket immunity under Section 230 of the Communications Decency Act of 1996, 47 U.S.C. § 230(c),(e). Again, however, Google overstates the protections the CDA provides. It is undoubtedly true that courts have interpreted Section 230 to protect service providers, including Google. See, e.g., Zeran v. America Online, Inc., 129 F.3d 327, 330 (4th Cir. 1997). Computer service providers like Google are protected from certain forms of liability where they publish information provided by another information content provider, see id.; 47 U.S.C. § 230(c)(1). The But the courts have also been clear that this immunity is not unlimited. The CDA “was not meant to create a lawless no-man’s-land on the Internet.” Fair Housing Council of San Fernando Valley v. Roommates.com, LLC, 521 F.3d 1157, 1164 (9th Cir. 2008) (en banc). The And the CDA does not shield Google from its conduct. Section 230(c) of the CDA provides no immunity to an internet service provider like Google when it, rather than a third party, is the “information content provider.” See id. at 1162; FTC v. Accusearch Inc., 570 F.3d 1187, 1197-1200 (10th Cir. 2009). That is, if the service provider is “responsible, in whole or in part, for the creation or development of” the offending content, 47 U.S.C. § 230(f)(3), its actions fall outside the protection of Section 230. As a result, courts have held that a service provider is not immune from suit where the provider itself creates or helps to develop, rather than merely publishes, the unlawful content. See, e.g., Roommates.com, 521 F.3d at 1168-69; Accusearch, 570 F.3d at 1198-99; Anthony v. Yahoo! Inc., 421 F. Supp. 2d 1257, 1262-63 (N.D. Cal. 2006). Where a search engine is “much more than a passive transmitter of information provided by others,” and instead “becomes the developer, at least in part, of that information,” Section 230 offers no protection. Roommates.com, 531 F.3d at 1166. For this reason, the CDA provides Google no immunity for its wrongdoing.[11] Google is not a mere publisher of third-party content when it suggests search terms through Autocomplete. Google authors the algorithm that generates the suggestions, and Google alters those suggestions based on the identity of the user and to ensure that the user is not directed to offensive content. Thus, Google is the developer of the content generated by Autocomplete. When Autocomplete steers users towards illegal content and websites, Google is responsible and outside Section 230’s protections. See Roommates.com, 521 F.3d at 1167. Similarly, where Google’s AdWords program assists criminals in optimizing their advertising campaigns, as Google conceded it did for illegal pharmacies in the NPA, it is an information content provider excluded from the protections of the CDA. Such advertising campaigns are not solely attributable to a third party because Google has itself created or developed, “in whole or in part,” 47 U.S.C. § 230(f)(3), the unlawful advertising campaign. Moreover, Google enjoys no protection under the CDA where it itself engages in conduct that is unlawful – regardless of who is the “publisher” of content. On YouTube, Google enters into contracts with the producers of YouTube videos to monetize illegal content and fund criminal activity. Google becomes a business partner, sharing advertising profits with criminals. Such aiding and abetting of criminal activity falls outside the immunities of Section 230 of the CDA. Courts have been quite clear that the CDA offers no protection to service providers that have themselves engaged in unlawful practices. See, e.g., Anthony, 421 F. Supp. 2d at 1263 (The CDA “does not absolve Yahoo! from liability for any accompanying misrepresentations” Yahoo! itself made); Mazur v. eBay, Inc., 2008 U.S. Dist. LEXIS 16561, at *28 (N.D. Cal. Mar. 4, 2008) (“The CDA does not immunize eBay for its own fraudulent misconduct.”); 800-JR Cigar, Inc. v. Goto.com, Inc., 437 F. Supp. 2d 273, 295 (D.N.J. 2006) (holding that the CDA cannot “shield entities from claims of fraud and abuse arising from their own pay-for-priority advertising business, rather than from the actions of third parties”). At its core, Google’s arguments under Section 230(c) are that it bears no responsibility for any criminal activity occurring on any of Google’s various platforms – regardless of the role Google has taken in creating, developing, encouraging, and profiting from that conduct. This But this is not about holding Google liable for merely being a conduit for the speech and actions of others. It is about holding Google to account for its own knowledge and actions – its facilitation of and profit from unlawful conduct, its own choices and actions in building its search and other algorithms, its promotion of particular unlawful websites through the Autocomplete feature that it created and wholly controls, and its business partnership with the producers of YouTube videos engaged in unlawful conduct. No entity – not even Google – is above the law. Sincerely yours, Jim Hood Attorney General cc: Jamie Gorelick The Honorable Jon Bruning The Honorable David Louie The Honorable Eric Holder ________________ [1] The search for “buy oxycodone” was conducted on August 29, 2013. [2] This is not about the free flow of expression protected by the First Amendment or about the challenge of previewing the trillions of web pages in the world. Those are strawpersons set up by Google’s Letter. Rather, this is about what Google knows, what it does, and what it refuses to do (even though Google concedes that it can). [3] Press Release, Attorney General Asks Colleagues to Issue Subpoenas in Google Investigation (June 18, 2013), http://agjimhood.com/index.php/press/releases/attorney_general_asks_colleagues_to_issue_subpoenas_in_google_investigation. [4] These searches were conducted on August 29, 2013. [5] See Digital Citizens Alliance, Google & YouTube and Evil Doers: Too Close for Comfort (June 2013), http://www.digitalcitizensalliance.org/cac/alliance/getobject.aspx?file=YouTube. [6] See id. [7] https://support.google.com/youtube/answer/94522?hl=en&ref_topic=1322133. [8] Id. [9] See Non-Prosecution Agreement between Google, Inc. and the U.S. Attorney’s Office for the District of Rhode Island (Aug. 2011), http://www.justice.gov/usao/ri/news/2011/august2011/Google%20Agreement.pdf. [10] This was one of the monetized videos specifically discussed in the Digital Citizens Alliance report. See Digital Citizens Alliance, supra at 9-10. [11] The CDA also offers no immunity for Google’s violations of federal criminal law, see 47 U.S.C. § 230(e)(1). It therefore cannot shield Google from liability for its facilitation of forgery of identification, counterfeiting, illegal drug sales, piracy, human trafficking, and other federal crimes.