The language of privacy and security, as articulated by Google’s statements on data protection, does not address what happens when you want your data to be deleted or forgotten. Indeed, Google suggests that when you delete data from an application, it is wiped from the Google servers:
Deleted Data
After a Google Apps user or Google Apps administrator deletes a message, account, user, or domain, and confirms deletion of that item (e.g., empties the Trash), the data in question is removed and no longer accessible from that user’s Google Apps interface.
The data is then deleted from Google’s active servers and replication servers. Pointers to the data on Google’s active and replication servers are removed. Dereferenced data will be overwritten with other customer data over time.21
But these explanations do not address the myriad ways that records are created and circulated through Google’s products and how we lose control over information about ourselves. Recently, Darlene Storm wrote an article for ComputerWorld citing researchers who purchased twenty mobile phones from Craigslist and eBay only to find thousands of photos, emails, and texts—including deleted messages through Facebook—after doing factory resets of their data.22 The most acute breaches of personal security were on Android smartphones, after using Google’s software to allegedly wipe them clean. Personal information at the level of device and infrastructure is not forgotten and can be circulated with ease.
The ways in which our human activities are recorded and stored are vast, and the value of social forgetfulness is not just good for individuals but is good for society. We should frame it as a public or social good:
A world in which there is no forgetfulness—a world in which everything one does is recorded and never forgotten—is not a world conducive to the development of democratic citizens. It is a world in which one must hesitate over every act because every act has permanence, may be recalled and come back to haunt one, so to speak. Of course, the opposite is equally true: A world in which individuals are not held accountable over time for the consequences of their actions will not produce the sense of responsibility that is just as necessary to a democratic society. Thus, achieving the appropriate degree of social forgetfulness is a complex balancing act, ever in tension between the need to hold accountable, and the need to grant a “fresh start.”23
Google’s position about forgetting has stood in stark contrast to previous conceptions of memory and forgetting, as Napoleon Xanthoulis of the Dickson Poon School of Law at King’s College London articulated in his important article theorizing the rights of individuals to control their data privacy as a fundamental human rights issues: a “right to cyber-oblivion.” He notes that Google’s chief privacy officer, Peter Fleisher, has argued against cyber-oblivion, or record wiping, as “an attempt to give people the right to wash away digital muck, or delete the embarrassing stuff.”24 Indeed, Google’s position has been that the recording of everything we do is a matter of the cultural record of humanity, “even if it’s painful.”25 Both Xanthoulis and Blanchette and Johnson argue that it is important that bad actors, violators of the public trust, and ill-intentioned public officials not necessarily be allowed to erase their deeds from the digital record. This has been Google’s general disposition toward erasures of information from its records. However, Google has begun to respond to pressures to change its algorithm. On August 10, 2012, Google announced on its blog that it would be pushing further down in its ranking websites with valid complaints about copyright infringement.26 Google suggested that this would help users find more credible and legitimate content from the web. This decision was met with much commendation from powerful media companies—many of which are Google’s advertising customers. These companies want to ensure that their copyrighted works are prioritized and that pirated works are not taking prominence in Google’s web results.
***
There are many troubling issues to contend with when our every action in the digital record is permanently retained or is retained for some duration so as to have a lasting impact on our personal lives. Privacy and identity ownership are constructed within a commercial web space such as Google, and Google controls the record. Subjects and publics are documented through Google’s algorithms, and displays of search results are decidedly opportunistic and profitable. While tremendous focus on “right to be forgotten” legislation is on control of records that are publicly visible on the web (e.g., websites, images, audio files, etc.), more attention needs to be paid to information that is collected and archived by Google that is not visible to the public. These records are conveyed by Google as necessary for its product development and for enhanced consumer experiences (see Google’s privacy policy). However, Google’s record keeping has its genesis in providing information shared across its networked services for its clients, which include U.S.-based national security agencies, as well as Google’s commercial partners. Increased attention must be paid to both the visible and invisible ways that identity information and records of activity can be archived through Internet infrastructures, buttressed by Google’s monopoly on information services in the United States. Inevitably, the power differentials between the record keepers, in this case a private company such as Google, and those who are recorded are insurmountable. Google’s power is only buttressed by its work on behalf of the U.S. government, which has outsourced its data collection and unconstitutional privacy invasions to the company.27
The goal of elevating these conversations is to recognize and name the neoliberal communication strategies used by Google to circumvent or suppress its record keeping of the public through surveillance, particularly in its privacy policies and responses to “right to be forgotten” public policy. Google’s control and circumvention of privacy and the right to be forgotten intensifies damage to vulnerable populations. As I have previously argued, Google is, at one moment, implicated in prioritizing predatory misrepresentations of people, such as algorithmically privileging sexualized information about women and girls, because it is profitable. In another moment, it is providing our records to third parties. While Google has consistently argued that “right to oblivion” laws are unfairly shifting the record of real-world human activity, which it believes the public has a right to know, recent leakages of requests for take-down notices were reported in the British media, showing that the nature of take-down requests is much more personal and relevant to everyday people, rather than public figures skirting responsibility to some alleged public interest.
On July 14, 2015, the Guardian reported that “less than 5% of nearly 220,000 individual requests made to Google to selectively remove links to online information concern criminals, politicians and high-profile public figures . . . with more than 95% of requests coming from everyday members of the public.”28 What is critical to this new revelation is that previously Google’s statements about the nature of “right to be forgotten” requests have been exaggerated or unknown because delisting information from its records has not been transparent, despite calls for information about the nature of the requests from over eighty academics in a letter authored by Ellen P. Goodman, a professor of law at Rutgers University, and Julia Powles, a researcher at the University of Cambridge Faculty of Law.29 In an open letter to Google, the scholars not only argue that the public has a right to have information taken out of Google’s search engine and all other engines subject to data protection rulings, but they also state,
Google and other search engines have been enlisted to make decisions about the proper balance between personal privacy and access to information. The vast majority of these decisions face no public scrutiny, though they shape public discourse. What’s more, the values at work in this process will/should inform information policy around the world. A fact-free debate about the RTBF [right to be forgotten] is in no one’s interest.30
Challenging content on the web under the auspices of the right to be forgotten must extend beyond the take down of personal information and beyond erasing the memory of pas
t acts from the web. The right to be forgotten must include the recognition of all forms of records that Google is archiving and sharing with third parties, both visible and invisible to the public.
The discussion about the right to be forgotten has largely lived in the frameworks of contesting neoliberal control and encroachments on social and public life organized around unstable notions of a public sphere. In the academics’ letter, calls for transparency about delisting requests point to the ways that ideologies of transparency privilege a kind of fact-based, information-oriented gathering of evidence to make clear and thoughtful decisions within the context of how privacy should operate within the records of Google. The questions about who controls the records of our social life and how they can be forgotten must move to the fore in the United States. They are explicitly tied to who can own identity markers and how we can reclaim them at both the individual and community level.
Librarians and information professionals are particularly implicated in these projects. In 2016, the librarian Tara Robertson wrote an important blog post to the profession about why all information should not be digitized and made available on the open web. Robertson’s point is that people share material, thoughts, and communications with each other in closed communities, as in the case of the digitization of On Our Backs, a lesbian porn publication that had a limited print run and circulated from 1984 to 2004, prior to the mainstreaming and commercialization of content on the web that we see today. People who participated in the publication did so before there was an Internet, before digitization would make the material public.31 Robertson raises the important ethical issues, as have many other researchers, about what should be digitized and put on the open web and what belongs to communities with shared values, to be shared within a community:
In talking to some queer pornographers, I’ve learned that some of their former models are now elementary school teachers, clergy, professors, child care workers, lawyers, mechanics, health care professionals, bus drivers and librarians. We live and work in a society that is homophobic and not sex positive. Librarians have an ethical obligation to steward this content with care for both the object and with care for the people involved in producing it.32
Figure 4.1. Call to librarians not to digitize sensitive information that was meant to be private, by Tara Robertson.
On Our Backs has an important history. It is regarded as the first lesbian erotica magazine to be run by women, and its title was a cheeky play on the name of a second-wave, and often antipornography, feminist newspaper named Off Our Backs. On Our Backs stood in the sex-positive margin for lesbians who were often pushed out of the mainstream feminist and gay liberation movements of the 1970s–1990s. What Robertson raises are the ethical considerations that arise when participants in marginalized communities are unable to participate in the decision making of having content they create circulate to a far wider, and outsider, audience. These are the kinds of issues facing information workers, from the digitization of indigenous knowledge from all corners of the earth that are not intended for mass public consumption, to individual representations that move beyond the control of the subject. We cannot ignore the long-term consequences of what it means to have everything subject to public scrutiny, out of context, out of control.
Ultimately, what I am calling for is increased regulation that is undergirded by research that shows the harmful effects of deep machine-learning algorithms, or artificial intelligence, on society. It is not just a matter of concern for Google, to be fair. These are complex issues that span a host of institutions and companies. From the heinous effects manifested from Dylann Roof’s searching on false concepts about African Americans that may have influenced his effort to spark a race war, to the ways in which information can exist online about people and communities that can be nearly impossible to correct, to the owning of identity by the highest bidder—public policy must address the many increasing problems that unregulated commercial search engines pose. In addition to public policy, we can reconceptualize the design of indexes of the web that might be managed by librarians and information institutions and workers to radically shift our ability to contextualize information. This could lead to significantly greater transparency, rather than continuing to make the neoliberal capitalist project of commercial search opaque.
5
The Future of Knowledge in the Public
Student protests on college campuses have led to calls for increased support of students of color, but one particular request became a matter of national policy that led to a threat to the Library of Congress’s budget in the summer of 2016. In February 2014, a coalition of students at Dartmouth College put forward “The Plan for Dartmouth’s Freedom Budget: Items for Transformative Justice at Dartmouth” (the “Freedom Plan”),1 which included a line item to “ban the use of ‘illegal aliens,’ ‘illegal immigrants,’ ‘wetback,’ and any racially charged term on Dartmouth-sanctioned programming materials and locations.” The plan also demanded that “the library search catalog system shall use undocumented instead of ‘illegal’ in reference to immigrants.” Lisa Peet, reporting for Library Journal, noted,
The replacement of the subject heading was the culmination of a two-year grassroots process that began when Melissa Padilla, class of 2016, first noticed what she felt were inappropriate search terms while researching a paper on undocumented students at Dartmouth’s Baker-Berry Library in 2013. While working with research and instruction services librarian Jill Baron, Padilla told LJ [Library Journal], she realized that nearly every article or book she looked at was categorized with the subject heading “Illegal aliens.”2
The Dartmouth College librarians became deeply engaged in petitioning the Library of Congress. According to Peet, “Baron, DeSantis, and research and instruction services librarian Amy Witzel proposed that the students gather documentation to prove that ‘Illegal aliens’ is not a preferred term, and to find evidence that better terms—such as ‘Undocumented immigrant,’ which was their initial suggestion for a replacement—were in common use. At that point news organizations such as the Associated Press, USA Today, ABC, the Chicago Tribune, and the LA Times had already committed not to use the term ‘Illegal’ to describe an individual.”3 Though unsuccessful in 2015, the librarians’ case to the Library of Congress had gained traction, and the librarian and professor Tina Gross at St. Cloud State University began organizing caucuses and committees in the American Libraries Association, including the subject analysis committee, social responsibilities round table, and REFORMA, which advocates for library services for Latinos and those who speak Spanish. Social media campaigns ensued, organized under the Twitter hashtags #DropTheWord and #NoHumanBeingIsIllegal.4 By March 29, 2016, Dartmouth College’s student-led organization the Coalition for Immigration Reform, Equality (CoFired) and DREAMers announced in a press release that after a two-year battle, in partnership with campus librarians and the American Libraries Association, “the Library of Congress will replace the term ‘illegal aliens’ with ‘noncitizens’ and ‘unauthorized immigrants’ in its subject headings.”5
“Illegal Alien” Revisited
The struggle over reclassifying undocumented immigrants was part of a long history of naming members of society as problem people. In many ways, this effort to eliminate “illegal alien” was similar to the ways that Jewish people were once classified by the Library of Congress as the “Jewish question,” later to be reclassified in 1984 as “Jews,” and Asian Americans were once classified as the “Yellow Peril.”6 Control over identity is political and often a matter of public policy. Almost as soon as the successful change was approved, House Republicans introduced HR 4926 on April 13, 2016, also known as the “Stopping Partisan Policy at the Library of Congress Act,” sponsored by Rep. Diane Black (R-TN). In essence, the bill threatened the Library’s budget, and Black suggested that the effort to change the Library of Congress Subject Headings (LCSH) was a matter of “caving to the whims of left-wing special interests and attempting t
o mask the grave threat that illegal immigration poses to our economy, our national security, and our sovereignty.”7
The battle over how people are conceptualized and represented is ongoing and extends beyond the boundaries of institutions such as the Library of Congress or corporations such as Alphabet, which owns and manages Google Search. Jonathan Furner, a professor of information studies at UCLA, suggests that information institutions and systems, which I argue extend from state-supported organizations such as the Library of Congress to the Internet, are participating in “legitimizing the ideology of dominant groups” to the detriment of people of color.8 His case study of the Dewey Decimal Classification (DDC) system, for example, underscores the problematic conceptualizations of race and culture and efforts to “deracialize” the library and classification schemes.9 Furner offers several strategies for thinking about how to address these issues, using critical race theory as the guiding theoretical and methodological model. I believe these strategies are of great value to thinking about the information studies issues at hand in this research:
• admission on the part of designers that bias in classification schemes exists, and indeed is an inevitable result of the ways in which they are currently structured;
• recognition that adherence to a policy of neutrality will contribute little to eradication of that bias and indeed can only extend its life; [and]
• construction, collection and analysis of narrative expressions of the feelings, thoughts, and beliefs of classification-scheme users who identify with particular racially-defined populations.10
Algorithms of Oppression Page 14