Dark Data

Background

This static visualization explores the evolution of Facebook’s Terms of Service. The first is from November 2007 and the second is from March 2019. It investigates the language used, and any changes made, frequently used terms, linked terms, the ambiguity of data handling and technological changes within the terms. It will focus on the following keywords data, products, privacy, and third-party.

Then (2007)

The language used in Facebook’s revised Terms of Service (TOS) dated November 15, 2007 ¹, differs considerably from its most recent iteration. A sizeable chunk of the Terms of Service is dedicated to user conduct and user content. The 2007 TOS has nearly twice the number of words (6400) than the 2019 TOS (3200). Noticeably missing from 2007’s version is Facebook’s use of user data for monetary gain. The onus and responsibility of the content are put on the Facebook user. The semantic content of the 2007 TOS seems to suggest that Facebook places more weight on a “community-based platform”, in the sense that Facebook is merely the facilitating medium between users. User content is treated as such: “When you post User Content to the Site, you authorize and direct us to make such copies thereof as we deem necessary in order to facilitate the posting and storage of the User Content on the Site.”

Third-party content on Facebook in 2007 was legally designated as the responsibility of the user. “Such Third Party Sites and Third Party Applications, Software or Content are not investigated, monitored or checked for accuracy, appropriateness, or completeness by us, and we are not responsible for any Third Party Sites accessed through the Site or any Third Party Applications, Software or Content posted on, available through or installed from the Site, including the content, accuracy, offensiveness, opinions, reliability, privacy practices or other policies of or contained in the Third Party Sites or the Third Party Applications, Software or Content.” There is no mention of personal user data being sold for profit. Interestingly, any content relating to data-mining and data-gathering is in relation to the user’s unauthorized use of data, such as the user selling data scraped from the platform itself and not relating to the platform’s mass data gathering for monetization.

The only mention of security relating to the use of personal data in the 2007 TOS again puts the onus on the Facebook user. “In consideration of your use of the Site, you agree to (a) provide accurate, current and complete information about you as may be prompted by any registration forms on the Site ("Registration Data"); (b) maintain the security of your password and identification; (c) maintain and promptly update the Registration Data, and any other information you provide to Company, to keep it accurate, current and complete…”. Considering the year, this is not a surprise. Many serious data breaches occurred after 2007.

The word privacy occurs 16 times in the 2007 Terms of Service, while the word data occurs 29 times.

Now (2019)

The language behind Facebook’s most recently revised 2019 Terms of Service has shifted to emphasize user data mainly within a revenue-generating model. In contrast to the ‘opt-in’ terms which characterized the 2007 TOS, the 2019 TOS is more of an ‘opt-out’ contract, but in fact without many options to do so.

Some of the more egregious and questionable terms in the most recent TOS include: “Permission to use your name, profile picture, and information about your actions with ads and sponsored content: You give us permission to use your name and profile picture and information about actions you have taken on Facebook next to or in connection with ads, offers, and other sponsored content that we display across our Products, without any compensation to you.” Even if the user no longer agrees with the terms, Facebook will keep/store your content for a “limited period of time in backup copies”, though it is unclear how long the content actually remains on their servers.

Language in the 2019 TOS has shifted from user-generated content owned by the user of said content to Facebook owning every aspect of the value chain. “

Advanced analytics tools such as artificial intelligence, machine learning systems, and AR are mentioned yet the language in the 2019 TOS which pertains or refers to these tools does not specify exactly how they make use of user data.

The word privacy occurs 3 times in the 2019 Terms of Service, while the word data occurs 13 times.

In April 2019, Facebook announced that it would be updating its policy to better reflect how it will be using its user data following pressure from the European Union[1]. “The actual content of the terms of service won't change, but the language very much will. Facebook agreed to make the changes, which will apply globally, by the end of June.”² The EU Commissioner for Justice Consumers and Gender Equality, Vera Jourová, said in a statement. "A company that wants to restore consumers trust after the Facebook/Cambridge Analytica scandal should not hide behind complicated, legalistic jargon on how it is making billions on people's data."

It will be interesting to see the newest iteration of Facebook’s Terms of Service as it hopefully steps away from convoluted language. □

*note this analysis looks at the Terms of Service and not links affiliated with the TOS.

Simone Betito is an MS candidate in Data Visualization. She loves a good baroque painting and has mixed feelings about pie charts.