Outside Google filter bubbles

This fall, I didn’t recognize the ways that a search history can be harmful. Even if a site means to be polite, it can still cause problems.

One problem is that you could end up in a filter bubble. Inside the bubble, you see things you already knew. You can enter an echo chamber where you hear what you already believe.

So, I decided: Let’s get out of my Google filter bubble. It wasn’t easy (which didn’t surprise me) but I got it done in Firefox.

I blocked all Google and YouTube cookies as well as their saved login information.

My goal was narrow. I didn’t want to improve privacy around Google. There are better tools for that. I just wanted to eliminate Google’s awareness of who I am. I did my best to not affect other services unnecessarily.

Some consequences that I know about:

  • You can’t read Gmail.
  • You can’t turn on SafeSearch in Google Images.
  • You can’t leave comments or thumb videos on YouTube.
  • You can’t use Watch Later in YouTube.
  • You have no history in YouTube.
  • Google Books lets you view extremely little of books (less than 5 pages).
  • You get the same ad links over and over.

I did some more investigation about the content of search results but I’ll explore that later.

Continue reading “Outside Google filter bubbles”

Power Imbalances

In situations where privacy problems appear, there is often an imbalance of power. One of the parties in the interaction can make decisions without the consent of the other. Additionally, the other doesn’t have the ability to counteract these decisions. They may not have sufficient influence. They may not know about the decisions. They may be excluded from having the right to affect the decisions.

To identify what areas of privacy I want to study, I believe having a list of privacy actors could be helpful to weed out things that are important to me and I’m interested in. I’ll group the actors into categories.

There are other agents affecting privacy, but they do not involve a strong power imbalance. Those aren’t included in this list. Some of entities here that have their powers affected by legislation such as FERPA and HIPAA. This includes Medical providers, financial institutions, intermediaries and educational institutions and any government agency.

This table primarily concerns Solove’s Information Collection, Information Processing and Information Dissemination privacy problems. It doesn’t address the Invasion privacy problems.

Agents that collect data that has economic value to themselves or may sell it

  • Financial institutions such as banks, lenders, investment agencies and debt collectors. They collect detailed financial information.
  • Online merchants. Their privacy policies are difficult to influence.
  • Brick and Mortar Merchants. They can capture information about purchases and payment services such as credit cards. They also don’t need to document their policies.
  • Insurance companies. They want information that helps them decide insurability.

Agencies that work behind the scenes yet they impact privacy rights

  • Technology infrastructure providers. Both internet service providers and service provides like cable companies. They can decide privacy policies that the individual cannot refuse.
  • Technology manufacturers. Once something is manufactured, whether it is an automobile, smart phone, computer or internet of things object, users doesn’t have the ability to change the information it controls.
  • Data warehouses. The collect information and individuals have no control over what is there, how accurate it is or what it will be used for.

Also financial institutions and brick and mortar merchants can work behind the scenes.

Agencies with legal right to acquire some personal knowledge

  • Government agencies such as the license branch and IRS have more control over the information than the information source.
  • Law enforcement. They gather information for criminal prosecution. They can also request information that you are not obligated to give, but it is difficult to say no. For example, they may ask to see your smart phone.
  • The judicial branch of the government. Criminal cases and civil disputes can include private information.
  • Educational institutions may accumulate many forms information. This information may be available to future employers, loan and scholarship providers and funding sources for the school itself.
  • Medical providers. The collect detailed protected health information.

Also financial institutions.

Agents that make privacy decisions that in practice have no recourse. These entities are distinguished in that the individual is aware that they are interacting with the agents.

  • Technology service providers. Social media sites like Facebook and Twitter, multimedia sites like YouTube, Vimeo and Soundcloud, search engines like Google, Bing and DuckDuckGo, information sources like Wikipedia, Reddit and StackExchange. Some are near monopolies and thus have more power than an individual who needs to use their services.
  • Library service providers. Entities that control services used by the library such as the ILS and information resources such as databases and specialized services can set their own policies that the library or its patrons may not influence.

Also online merchants.

Agents that can affect privacy through surveillance

  • Employers. They can capture information about employee activities.
  • Intermediaries like the telephone network that have specific legal responsibilities but may also be used to gather information about transactions they assist.

Also law enforcement, government agencies, and the judicial branch of the government.

Data warehouse are not in this category because although they may enable surveillance, they are normally not the agents performing the surveillance

Agents that may find legal protection when they threaten privacy

  • Media outlets, publishers and online bloggers or similar publishers. They can disclose private materials without the individual’s consent or easy recourse.

Privacy and Self-Censorship

It’s interesting when the domain of two classes overlaps simultaneously. The subject of privacy and self-censorship is one of those right now.

A boy held back by several hands
The current assignment in IUPUI’s S502, Acquisitions and Management of Knowledge and Information explores censorship. Neil Richards’ Intellectual Privacy addresses the need to allow one to read freely and avoid self-censorship. The privacy interests in Richards’ book explain that observation or surveillance of ones reading materials can have a chilling effect on what information a person reads and accesses. One harm of that is that it may prevent the development of informed citizens in a democracy.

502 addresses the issues of censorship in a couple of narrow contexts. One form of censorship prevents access to objectionable materials. The other is self-censorship. To short-circuit possible challenges to a resource, the librarian may not purchase materials in an act of self-censorship. The class resources emphasize the value of librarians to pushing back against censorship attempts. There readings also mention self-censorship by authors who change their texts to improve a work’s palatability.

My observation is that the issue of censorship is broader than what 502 discusses. The self-censorship by authors is accurate. Publishers, content and copy editors, translators as well as authors can all implement self-censorship. Anyone in the food chain between the originator of an idea and the consumer is a vector for self-censorship.

Countries with authoritarian governments don’t try to disguise their censorship. They also use it to demand self-censorship. Post-publication censorship can apply pressure for publishers and news sources to self-censor. In the United States, this form of self-censorship is discreet. For example, authors don’t have to show that they removed profanity or didn’t write the article because of any controversies.

The class notes discussed self-censorship by librarians and booksellers. The issue of self-censorship from Intellectual Privacy discussed wasn’t as important to 502.

When an agency might monitor what a reader buys, rents or borrows, it isn’t necessary for there to be explicit restrictions. One might avoid owning a paper copy of with Fifty Shades of Grey that Richards explains in the chapter “Reading” that starts on p. 123. Censor’s attacks may inhibit readers to borrow books that they had attacked. One may not go to a newspaper web site published from Damascus, Tehran or by the Palestinians.

Richards points out a paradox. Ones reading material may seem private, for example by buying Fifty Shades of Grey as an e-book. However, eBook vendors have records of what you purchase, compromising the privacy and anonymity. The vendors protect that information out of economic self-interest, not the privacy needs of users. For example, in Ashcroft vs. ACLU, Google refused to release user search data with the justification that that release would reveal trade secrets. They did not advocate that the privacy of users was also important, just that the goodwill of users requires confidentiality.

An architectural diagram of the Panopticon prison
The Panopticon
The interaction between privacy and self-censorship comes from how one behaves when observed. The idea of the Panopticon prison invented by Jeremy Bentham is a metaphor in other venues. In the Panopticon, prisoners knew they could be watched at any time. They also did not know when they were not being watched. The risk of discovery forced the prisoners to act according to prison rules at all times. Similarly, awareness of possible access to internet history can lead to one to limit ones activities.

News reports mention internet access as incriminating in some criminal cases. The criminal complaint against the terror suspect Akram Musleh from Indiana included information about his search history [1]. However, that document contains much other incriminating information. His reading habits were secondary to other parts of the investigation. Other news reports sometimes mention search histories of narcotics suspects during the processing of their case.

This topic reveals the balancing needed to protect individual privacy. On one hand, individuals do not wish to be discovered searching for sensitive matters, but on the other, protecting that information can hide criminal activities and impede their prosecution. A group working to replace an oppressive government is criminal from within the country. However, observers from outside the country may encourage and try to protect their work.
The delineation of where public interests override individual needs is different in different points in history and varied political structures.

My inclination is to side toward allowing individual privacy to prevail. In only a limited set of circumstances is a history list relevant to building and prosecuting a criminal case. For example, all of a pedophile’s web activities would be directly relevant to the case. But, generally, reading or search records is supplementary information and restricting the use of the reading history will not derail a criminal investigation.

On the other hand, taking the use of online records even further, I can imagine a world where the search histories of government officials is just as relevant for public interest as their email records. For example, the web pages that the governor visits might contain nuggets of information that is revealing to his or her evaluation as a worthy candidate for re-election. In litigation related to patents, industrial espionage or trade secret disclosure, search and browser records may also be relevant to the case. The activity of the staff of a social media site might reveal that their service stole a feature from a minor player in the Web 2.0 biosphere.

Perhaps these last ideas seem alien only because people haven’t evaluated their potential. The problem is definitely not from any technical considerations. The use of these records could have a chilling effect on anyone in a position of authority’s effectiveness. As Richard’s emphasis, intellectual privacy in areas like this is valuable regardless of the technological possibilities.

[1] https://cchs.gwu.edu/sites/cchs.gwu.edu/files/downloads/Musleh%20Complaint.pdf
Original image: Censo****p. By NICHXAV. [Image license]
Original image: Panopticon. By 张 磊 [Image license]