Privacy Self Regulation: A Decade of Disappointment By Chris Jay Hoofnagle March 4, 2005
A hi-resolution version report is available in PDF (2.5 MB).
[Front Cover: Lists of personal information for sale from website registrations.][Inside Front Cover: Letter forwarded to EPIC explaining that an individual has no rights in personal information held by the company, Locateplus.com.]
Summary
The Federal Trade Commission (FTC) is capable of creating reasonable and effective privacy protections for American consumers. There is no better example of this than the Telemarketing Do-Not-Call Registry. The Registry, which was created and is now run by the FTC, makes it easy for individuals to opt-out of unwanted telemarketing. Now, more than 80 million numbers now no longer ring at the dinner hour.
Prior to the creation of the Registry, the telemarketing industry created self-regulatory protections that were largely useless. One had to write a letter to opt out of telemarketing, or pay to opt out by giving their credit card number to the Direct Marketing Association (DMA). The industry's self-regulatory efforts didn't even cover all telemarketers-only those that were members of the DMA. At its peak, the self-regulatory opt-out system had less then 5 million enrollments.
FTC's success in the telemarketing field demonstrates that it can protect Americans' privacy effectively and fairly. However, telemarketing was a 20th century problem. This report argues that it is time for the agency to move into the 21st century. It is time for the agency to apply the principles of telemarketing privacy regulation into the online world.
The FTC can protect privacy better than the industry can with self-regulation. We now have ten years of experience with privacy self-regulation online, and the evidence points to a sustained failure of business to provide reasonable privacy protections.
New tracking technologies exist that individuals are unaware of, and old tracking technologies continue to be employed. Some companies deliberately obfuscate their practices so that consumers remain in the dark. Spyware has developed and flourished under self-regulation. Emerging technologies represent serious threats to privacy and are not addressed by self-regulation or law.
Self regulation has failed to produce easy to use anonymous payment mechanisms.
And finally, the worst identification and tracking policies from the online world are finding their way into the offline world. In other words, the lack of protection for privacy online not only has resulted in a more invasive web environment, but has also started to drag down the practices of ordinary, offline retailers.
EPIC calls upon the Federal Trade Commission and Congress to seriously reconsider its faith in self-regulatory privacy approaches. They have led to a decade of disappointment; one where Congress has been stalled and the public anesthetized, as privacy practices steadily worsened. We call on the government to create a floor of standards for protection of personal information based on Fair Information Practices.
I. The FTC Registry Is Better Than Market Alternatives
The Federal Trade Commission's (FTC) Telemarketing Do-Not-Call Registry was a stunning privacy success. Americans enrolled 10 million numbers in the Registry in its first day of operation. Now, the phone has stopped ringing on the more than 60 million numbers that were enrolled by the public. The nuisance of telemarketing will now be a thing of the past. Those who wish to receive telemarketing may still do so, but others have an easy option to preserve the dinner hour from interruption.
When one analyzes the decisions made by the FTC, it reveals that the agency took steps to effect consumers' desires. The FTC publicized the existence of the Registry and gave it a simple name and URL on the Internet. The FTC allowed people to enroll free by telephone or by the Internet. The FTC minimized "authentication" burdens. That is, the FTC made it easy for people to enroll by not requiring the consumer to jump through unnecessary hoops. Some from the industry suggested that only the line subscriber-not even a spouse or roommate-could enroll.
The Do-Not-Call Registry was a success because the FTC took the opposite approach from the self-regulatory system created by the Direct Marketing Association (DMA). In every respect, the FTC ensured that the Registry would be easy to use and fair, while the DMA's opt-out mechanism was difficult to use and relatively unknown.
For starters, the DMA's system only applied to the industry association's members. Telemarketers who had not joined the group were not bound to comply with consumers' desire to opt-out. The FTC's approach applied to a much broader group of telemarketers.
Second, the DMA's list was named the "Telephone Preference Service." The name and acronym, "TPS," had no meaning to the public. To some, it could mean a list of people who preferred to be telemarketed. The FTC approach, on the other hand, was sensibly named and assigned a easy to remember URL,
Third, the DMA's list required the consumer to actually write a letter for free enrollment. To enroll online, the consumer had to pay a fee and give their credit card number to the DMA. The FTC's approach allows free Internet, mail, and telephone enrollment.
The FTC's Registry is universal, free, and easy to use. Individuals could enroll online or by phone. The DMA's only applied to its members, cost money to enroll online, and was difficult to find. It's no wonder why the DMA's list only had 5 million enrollments, while the FTC's has more than 80 million.
These forces combined to make the DMA's market approach to telemarketing ineffective. The numbers speak for themselves. USA Today commented in 2002 that: "In 17 years, just 4.8 million consumers have signed up with the DMA's do-not-call list. By contrast, just five states -- New York, Kentucky, Indiana, Florida and Missouri -- have signed up roughly the same number in far less time."[i]
Today's self-regulatory approaches to Internet privacy are much like the failed ones employed by the DMA for telemarketing. They are difficult to use, confusing, and often offer no real protection at all. This report details the current state of privacy on the Internet, and illustrates the myriad ways in which threats to privacy are becoming ever more grave, as new technologies are developed, new practices become commonplace, and companies are not held accountable for disregarding privacy risks. Collection of personal information on the Internet runs rampant, both through direct and indirect means, both in the open and in secret. It is imperative that the FTC act now to correct these market failures. The FTC effectively and fairly corrected the failures of a 20th century nuisance-telemarketing. It is time for the agency to move into the 21st century and correct the failures of self-regulation to meaningfully protect Internet privacy.
II. Ten Years of Self-Regulation and Still No Privacy In Sight
EPIC has completed three Surfer Beware reports assessing the state of privacy on the Internet. "Surfer Beware I: Personal Privacy and the Internet," a 1997 report, reviewed privacy practices of 100 of the most frequently visited web sites on the Internet. It checked for collection of personal information, establishment of privacy policies, cookie usage, and anonymous browsing. The inquiry found that few sites had easily accessible privacy policies, and none of these policies met basic standards for privacy protection. However, at that time, most of the sites surveyed allowed users to access web content and services without disclosing any personal data. The report ended with a recommendation of continuing support for anonymity and the development of both good privacy policies and practices.
In 1998, EPIC produced "Surfer Beware II: Notice Is Not Enough," a report based on a survey of the privacy practices of 76 new members of the Direct Marketing Association ("DMA"), a proponent of self-regulation of privacy protection. The DMA released guidelines in
1997 that would require all future members of the DMA to publicize privacy policies and provide an opt-out capability for information sharing. Of the 76 new members surveyed, only 40 had web sites, and only 8 of these sites had policies satisfying the DMA's requirements. The report concluded that DMA's self-regulation efforts were not effective.The 1999 report "Surfer Beware III: Privacy Policies without Privacy Protection" assessed the privacy practices of the 100 most popular shopping web sites on the Internet. It examined whether these sites complied with common accepted privacy principles, used profile-based advertising, and employed cookies. The survey determined that 18 of the sites had no privacy policy displayed, 35 of the sites used profile-based advertising, and 86 of the sites used cookies. None of the companies adequately addressed Fair Information Practices, commonly-accepted responsibilities covering collection, access to, and control over personal information. Surfer Beware III concluded that current practices of the online shopping industry provided little meaningful privacy protection for consumers.
The Federal Trade Commission ("FTC") has given self-regulation a decade to produce reasonable privacy protections online. The FTC first visited online privacy in 1995, and with minor fluctuations since then, has adopted a policy that embraces the idea that self-regulation is "the least intrusive and most efficient means to ensure fair information practices online, given the rapidly evolving nature of the Internet and computer technology."[ii] It certainly is the least intrusive approach for companies exploiting personal information, but it has not efficiently ensured Fair Information Practices. Of the five Fair Information Practices[iii] endorsed by the FTC-notice, choice, access, security, and accountability-only notice can be said to be present as a result of privacy statements.
The first fluctuation in the FTC's commitment to self-regulation occurred in 1998, after the agency's survey of online practices showed that the lowest level of protection for consumer, notice of privacy practices, was not widely implemented. In a survey of 1400 web sites conducted by the Commission, 92% of the commercial sites collected personal information but only 14% had privacy notices. Of the commercial sites, only 2% had a "comprehensive" privacy policy.[iv] In reaction to these findings, the FTC was "still hopeful" that industry efforts would produce adequate privacy protections.[v] At the time, Chairman Pitofsky recommended that Congress pass legislation if self-regulation failed to produce significant progress.[vi]
A year later in testimony to Congress, the FTC renewed its faith in self-regulation, noting that many web sites had adopted privacy policies. But protections beyond mere disclosure of practices lagged behind. Only a small number of surveyed sites had incorporated choice, access, and security into their practices. No meaningful avenue for enforcement existed at all. Commissioner Sheila Anthony concurred with the report's findings but dissented from its recommendations, noting, "industry progress has been far too slow since the Commission first began encouraging the adoption of voluntary fair information practices in 1996. Notice, while an essential first step, is not enough if the privacy practices themselves are toothless. I believe that the time may be right for federal legislation to establish at least baseline minimum standards."
"Notice, while an essential first step, is not enough if the privacy practices themselves are toothless."
In 2000, a 3-2 majority of the FTC formally recommended that Congress adopt legislation requiring commercial web sites and network advertising companies to comply with Fair Information Practices.[vii] However, a year later with the appointment of a new FTC Chairman, the FTC embraced self-regulation again. Chairman Muris decided to focus the Commission's attention on enforcing existing laws rather than create new legislative protections for online privacy.[viii] Chairman Muris indeed has expanded privacy protections through the creation of a do-not-call list and with application of the agency's powers to prevent unfair and deceptive trade practices.
The overall effect of the FTC's approach has been to delay the adoption of substantive legal protection for privacy. The adherence to self-regulatory approaches, such as the Network Advertising Initiative that legitimized third-party Internet tracking and the Individual References Service Group principles that concerned sale of SSNs, allowed businesses to continue using personal information while not providing any meaningful privacy protection. Ten years later, online collection of information is more pervasive, more invasive, and just as unaccountable as ever-and increasingly, the public is anesthetized to it.
It doesn't have to be this way. The FTC has been effective in protecting privacy when dealing with 20th century nuisances. It's time for the FTC to apply the lessons from telemarketing and other efforts to address the 21st century problem of Internet privacy.
III. Today's Tracking Methods Are More Pervasive and Invasive
Seven years ago, EPIC's report Surfer Beware I reviewed the status of Internet users' privacy rights and protections on the 100 most frequently visited web sites. The report was concerned primarily with the solicitation, collection, use, and protection of personal information obtained either from user-input forms or cookies.
Today, there are many more methods through which users can be tracked, profiled, and monitored in the online world. Cookie technology has matured-cookies are widespread and new uses have been developed. Entirely new technologies have emerged as well, some of which are all but unknown to consumers. Few of these methods are regulated, either internally by industry or externally by government. Without privacy legislation to protect Internet users from improper use of the information collected on the web, companies are unlikely to voluntarily cease privacy-invasive practices.
Cookies
Surfer Beware I discussed an Internet tracking technology over which there was "a great deal of controversy"-cookies. It found that about a quarter of the most frequently visited web sites used cookies. Today, many websites use cookies for one reason or another. In addition, there are several new wrinkles in the use of this tracking technology.
Third Party Cookies
Today, websites that a user explicitly visits are not the only entities which place cookies in your web browser-many web sites contain advertising served by outside commercial providers, and these providers may also send a cookie to your browser. These are known as "third party cookies." Some web browsers, such as Firefox allow users to block third party cookies.
Many web pages today have arrangements with third party ad servers that serve advertisements to their pages. For example, the MSN Privacy Statement lists two dozen third party ad networks that may place cookies in a user's browser.[ix]
Privacy policies (such as MSN's) tend to frame these third party cookies as a benefit to the user, allowing advertisers to "deliver targeted advertisements that they believe will be of most interest to you."
Persistent Cookies
A persistent cookie is one that remains on a user's computer after she has quit the browser. These cookies can be used to set and remember a user's web site preferences, settings, and passwords from one browser session to the next, but can also be used for tracking and monitoring purposes. A troubling recent trend is to design these cookies to remain not just for many browser sessions, but for many years. Google's search cookie, for example, will not expire until January 17, 2038. This kind of long range tracking of users raises significant privacy risks.
Web Bugs
A web bug is a graphic on a web page that allows tracking and monitoring of visitors to that page. Web bugs are usually invisible, "clear" images only 1-by-1 pixel in size. They are capable of transmitting, back to the bug's originating server your Internet Protocol ("IP") address, the page you visited, the time you visited, browser information, and information from existing cookies in the browser.
For market approaches to work, consumers must grasp both technology and practices. But in a Pew Internet Report, 56% surveyed couldn't identify a cookie.[x]
Web bugs are sometimes used for the innocuous purpose of counting how many times a particular page is viewed and gathering statistics about browser usage and web site usage. There are, however, much more invasive uses, such as compiling a detailed web-browsing profile of a particular user.
Web bugs are designed specifically to be secret and invisible. Many Internet users today are aware of cookies, and may perceive them from the appearance of visible advertisements. There are also tools to manage cookies. Web bugs, however, can transmit information and set cookies even when there is no telltale banner advertisement on the website tipping off a user that information might be collected about them. Furthermore, just one "allowed" cookie from an ad network opens the door for all web bugs within that network to collect browsing information about that user. With companies such as DoubleClick, providing advertising to countless web sites, this risk is significant. For instance, if a user with a DoubleClick cookie in their browser loads a web page with a DoubleClick web bug on it, that bug can grab the identifying information in the cookie and transmit it back to the server along with the other information collected by the bug.
Google's Gmail Content Extraction
On April 1, 2004, Google announced the launch of their new Gmail service. Gmail is a web-based e-mail service offering one-gigabyte of e-mail storage to users. Gmail is supported by advertisers who buy keywords, much like the Google search engine's AdWords advertising program, which lead to targeted advertisements displayed alongside an e-mail message in a Gmail user's inbox. Gmail uses "content extraction" (a term from Google's patents) on all e-mails sent to and from a Gmail account in order to target the advertising to the user.
"If Google ogles your e-mail, will Ashcroft be far behind?"[xi]
Many privacy advocates hold the position that the Gmail service violates the privacy rights of both Gmail users and non-subscribers. Non-subscribers who e-mail a Gmail user have "content extraction" performed on their e-mail even though they have not consented to have their communications monitored, nor may they even be aware that their communications are being analyzed.
This is a significant development in Internet tracking technology because it is one of the first with the capacity and the structure to monitor and record not just transactional data and personal information, but the content of private communications.
Spyware
Spyware and adware are extremely invasive and annoying technologies that have flourished in the self-regulatory world of Internet privacy. Both can be broadly described as pieces of software placed on a user's computer by a third party that perform unwanted functions. Spyware and adware collect information about the user, sometimes in complete secrecy without the knowledge of the user. Some programs display pop-up ads on the user's monitor, while others track and record everything the user does online. Information is sometimes collected by the programs for the sole purpose of sending that data back to an advertiser, and other times used to immediately serve pop-up ads to the user. Users often inadvertently download and install spyware and adware along with other desired computer programs, most commonly file-sharing applications. McAfee, an Internet security firm that sells popular virus protection and other personal computer security programs, reported more than 2.5 million "potentially unwanted programs" on its customers' computers, as of March 2004.[xii]
[TELECOM Digest Editor's Note: In the next issue of the Digest, we will begin with Part IV of this essay, discussing even more nefarious schemes to invade your privacy getting started. PAT