Privacy and libraries
3 likes2,372 views
This document discusses privacy issues related to libraries. It begins by distinguishing between security problems that inadvertently violate privacy and legal uses of information that can still violate privacy. It explores definitions of privacy, why privacy matters for intellectual freedom and concerns over vulnerable individuals. The document examines how freedom to read relates to privacy historically and today. It reviews potential privacy laws and policies, why privacy problems occur, and known privacy issues regarding public records, reidentification, email, commercial privacy violations on the web and examples related to Facebook, ebooks, Amazon and social media buttons.
Privacy and libraries
- 1. Privacy and libraries LIS 644 Dorothea Salo
- 2. Privacy vs. security • A security problem can violate privacy, sure. • But the violation is inadvertent! And often involves some illegality! • We’re not talking about that today; we’ve done so already. • We’re talking about perfectly legal (usually) uses of information that still (potentially or actually) violate privacy. • This is a minefield. I don’t have all the answers. We MUST still ask the questions.
- 3. What is privacy, really? • “Exposure of personal information” is too easy an answer. • Exposure of what to whom, exactly? • danah boyd: “respecting context” • Consider your social circles. You have several of them. What happens to them online? • So privacy is partly the ability to practice “selective disclosure” (another boyd-ism). • Privacy is also trust that those we interact with will not betray us. • What’s betrayal in a library context, and how do libraries avoid it?
- 4. Why do we care? • Privacy of information use is a cornerstone of intellectual freedom. • ALA Code of Ethics: • III. We protect each library user's right to privacy and confidentiality with respect to information sought or received and resources consulted, borrowed, acquired or transmitted. • Additional concerns include: • Privacy of research subjects in collected data • Privacy of living individuals mentioned in archival materials • Privacy of confidential business records • Privacy for especially vulnerable individuals
- 5. What is “freedom to read?” • Historically: preventing Big Brother from watching patrons’ checkout histories • Remember the Patriot Act? How ‘bout that NSA? • What happens to circ records nowadays? How did moving circ records digital help privacy? • Today: is Big Brother the only threat? • Who else wants information, and will pay and/or bug to get it? • Historically: censorship and book banning • Today: is that all that keeps good information out of the hands of patrons?
- 6. Do we have privacy laws? • Not in the US. Not really. • We have libel and slander laws, and laws about children’s information online (remember those?). Limited “invasion of physical space” laws. • Consumer-privacy laws have been introduced in Congress. No joy... yet. • We often have laws/ordinances around library patron information. • And even in the absence of law we SHOULD (and usually do) have policy. • KNOW THE LAW AND POLICY where you are. • It’s different in Canada and Europe. • Canada has even taken Facebook to the cleaners once or twice.
- 7. Why do privacy problems happen? • Monetary benefit • Accident or ineptitude (see thedailywtf.com) • Privilege and associated thoughtlessness • Google Buzz: testing with its own (white, male, healthy, wealthy, educated) engineers and no one else. • Collapse of realspace social boundaries on the web • This is a service-design problem! Human beings manage just fine in realspace. Our online tools don’t give us the affordances we need to replicate that success online. • Web is more than the sum of its parts. • Most scarily: DESIRE TO OFFER GOOD SERVICE, e.g. recommender engines.
- 8. Libraries and privacy • It’s not as simple as “always protect patron privacy!” • What about libraries and social media? Will we wipe librarians off the Web? Really? • What about user studies for service improvement? • What about patron communities that form around our materials and services? • What about users who WANT to share what they read, watch, and listen to? Will we shake our paternalistic fingers at them and tell them no? • What about digitization projects? Research? • “How” is getting harder to figure out too.
- 10. Law hinders privacy research • DMCA! • If there’s information about you on a machine you own (or even one you don’t), and the only way you can find out about it is to hack the machine... • These lawsuits have happened. • Often they go away very quickly as the vocal tech community shames the plaintiff. • But Ed Felten of Princeton has been in and out of court so many times...
- 11. Public records • Back in the day, if you wanted a public record, you went to a physical building, combed through file cabinets, and paid for the privilege. • This is a variant on security-by-obscurity. • Now many public records are online. Easy discovery and easy access make them A LOT MORE PUBLIC. • How should we, as citizens, respond? • As records managers/archivists/librarians, how should we educate, train, and refer?
- 12. An example
- 13. An example
- 14. Email • So how ‘bout that Petraeus guy? • Email is only very loosely legally protected at present. • Larger point: we have the privacy protections we do because they are enshrined in law, not because they’re societal norms. They’re pretty clearly not. • Email is sent in the clear unless you take encryption precautions. • Even then it may be readable if your inbox is hacked. • Your employer owns its email systems and email sent on them. Behave accordingly. • Students: using non-university email may bypass FERPA protections.
- 15. What is “reidentification” or “de-anonymization”? • Imagine this scenario: • One website has your name, age, zipcode, and gender. • Another has your age, gender, zipcode, pseudonym, and dubious or sensitive taste in entertainment. • If the info from both sites can be collated, you can be pegged to your taste. • And your pseudonym just got exposed. Hope you weren’t using it anywhere else... • We aren’t as unique as we think! • “Anonymizing” data doesn’t fix this. • We can be identified by our attributes, friends, and behavior almost as easily as by regular identifiers. • What price public records NOW?
- 16. Reidentification horror scenarios • Health information • Wouldn’t your insurance company like to know...? • Becoming a major issue in health research! • “Character witnessing” • Are you an atheist? A gamer (this came up in a 2012 political campaign)? GLBTQ and not out? A person of color who’s passing? A woman in IT? A whistleblower? • A target for harm • Physical, legal, financial, employment, mental/emotional (bullying) • Where could library-patron information figure in to this? Archives informants?
- 17. Commercial privacy violation on the web
- 18. What information does the web collect about us? • “Personal information” • Including health information, demography. • Financial information • Information about our habits • Purchasing habits • Entertainment habits (including, yes, reading habits) • Search habits • Information about our physical location • through IP addresses or through web services like Foursquare • Information about our social lives • And then it correlates as much of this as it can!
- 19. How is this information collected? • Through server and search logs (IP addresses) • Through sign-ins • some of which are “real name required” • Geolocation of our gadgetry • Browser “fingerprinting” • Which version, with what add-ons, on which OS... unique! • Human error (and exploitation thereof ) • Through observation of our behavior on individual websites and across websites • Cookies, Flash cookies, “web bugs.” Worst case: “keyloggers.” • Our online associates’ behavior • Which we obviously don’t control! • How much of this are we actually aware of? How much do sites disclose? Let us control?
- 20. Effects • ... on citizenship • ... on open discourse • ... on vulnerable populations • ... on markets • is privacy-endangerment a winner-take-all market? • what about online redlining?
- 21. Examples
- 22. Privacy and ebooks • Ebook vendors, unlike libraries, do not necessarily purge records of what you read. • You are entirely at their mercy as far as who they share those records with and what they do with them. • Are they collecting info from library patrons too? Unclear! • Because of this and DRM, they can also take away what you want to read. • And then there’s what you search for, or look at, but don’t read. • What do we do about this? What should we do?
- 23. Facebook has sold... • • • • Your phone number Information about your purchases Information about your social network Information about Facebook campaigns you’ve participated in • Information about what you’ve “liked.” • While refusing to let you opt out of the sale of this information. • Your likeness, for advertisers to use on your friends. • Google ain’t much better, and is getting worse.
- 24. Others have tried to use Facebook to... • Screen employees • including by requiring applicants to hand over Facebook passwords! • (To Facebook’s credit, it actually fought this one.) • Perform background checks (for employment or other reasons) • Do social-science research, sans informed consent • At Harvard, some researchers made their RAs hand over their Facebook passwords so they could see friendslocked material. • How are you feeling about your Facebook?
- 25. Guess what? • Facebook has sold MY information too, and I refuse to use Facebook! • Look up “shadow profiles” sometime. • If you delete your account, Facebook keeps and continues to sell your information. • Facebook may or may not actually delete photos when you delete them. • Guess why I don’t use Facebook? • Should libraries? Conflict between privacy ethics and “go where the patrons are.”
- 26. “Like” buttons • When you log into Facebook, Facebook knows you visited any page with a “Like” button on it, even if you do not click Like. • Facebook has also been caught tracking this on logged-out users. They claim they’ve stopped. • If your library puts Like buttons on catalog pages... (you do the math) • Not just a Facebook issue, by the way. • Social-media truism: • “If you are not paying for it, you’re not the customer; you’re the product being sold.” —blue_beetle on MetaFilter
- 27. Amazon • OverDrive signs a deal with Amazon to lend Kindle ebooks through libraries. • To do this, patrons have to tell Amazon their Kindle identifier, just as though they were buying the book. • Amazon sends “hi, your loan is ending, how about buying the book?” messages to patrons. • And is, as far as anybody knows, keeping information about who checked out what.
- 28. Try it yourself: JSTOR • JSTOR “Register and Read” program • Give non-affiliated scholars/interested public unpaid access to JSTOR, in return for a signup that ties reading to the signup’s email address. • Let’s look at their privacy policy. • http://www.jstor.org/page/info/about/policies/privacy.jsp • What info are they collecting? Reidentification risk? • What risks might there be to program participants with respect to what they read? • What do they say they can do with it? • How is this different from standard library policies, practices, and legal protections?
- 29. Try it yourself: JSTOR • Real reason to worry: Swartz case. • Is loss of privacy an unintended side effect of library disintermediation/disruption? • If so, what do we do? Without sounding like a bunch of luddite worrywart Trithemiuses just out to protect our own jobs?
- 30. Privacy in archives • Boston College IRA case • Oral histories collected from Northern Irish people who fought for IRA • Archivists promised informants not to release until after those informants died. • UK authorities: “Fork it over, archivists.” • Lawsuits flew! • What would you do? • You need to decide this. Before something similar happens to you.
- 31. What now?
- 32. What people want • Control of which pieces of data they share. • Choice about how their data will be used. • Commitment that their personal data (i.e., email address, phone number) won't be passed on to third parties. • Compensation: Consumers also want a reason to share data, and to understand how they will benefit. • (via http://www.mediapost.com/publications/article/161410/consumerswilling-to-share-data-but-at-a-price.html) • Can we do this in libraries? How?
- 33. More suggestions • Don’t collect data you don’t need. • And throw away data once it’s no longer of use. • This includes computer logs! (IM chat ref, anyone?) • Think outside your own demographic box. • As Google seems to have so much trouble doing... • Be transparent. • Be activist. We have a bully pulpit! • PAY ATTENTION to the security and privacy of library IT infrastructure. • This EMPHATICALLY includes the ramifications of thirdparty IT such as “like” buttons. • It also includes contracts with content providers. A privacy review should be an intrinsic part of collection development.
- 34. Rule of thumb? • In the absence of a warrant or subpoena, don’t keep or disclose information about the behavior of identifiable patrons until the patron has not only consented, but ASKED YOU to retain or disclose the information. • AND MAYBE NOT EVEN THEN. • We know people make poor choices here!
- 35. Protecting digital privacy • My suggestions: encryption, deletion, awareness. • Encryption is where it’s at, folks. It’s not perfect, but it’s the best we’ve got. • Delete digital records. As often as possible. Perhaps oftener. (Sorry, records managers and digital archivists! Privacy comes first!) • Try to be aware of when your data are being collected. Websites like tosdr.org (and the associated browser plugins) help!
- 36. Example: cloud storage • Cloud storage services almost all encrypt data at some point. • Google Drive, not so much. Just so you know. • Important questions: who holds the key, and when are the data locked up? • Dropbox, Box (for now): They hold the key. This means they can rat you out, snoop, etc. Also means that data travel in the clear, and are vulnerable to packet-sniffing! NOT SECURE. • SpiderOak: YOU hold the key, and encryption happens on YOUR machine, before data move over the network. SpiderOak doesn’t even see your data unencrypted, can’t decrypt it. Secure, but don’t lose passwords!
- 37. Example: protecting your web surfing from marketers • Remember the stuff I discussed last week with respect to browser security? It can help protect your privacy as well. • • • • • On an untrusted network, use a VPN to prevent packet-sniffing. Do not let your browser accept third-party cookies. Use adblocking, tracking-blocking browser add-ins liberally. Grab the “HTTPS Everywhere” browser add-in from the EFF. Turn on the “Do Not Track” setting in your browser; it doesn’t do much, but it does something at least. • Serious question: which of these should we install on patron computers? • Or is that too paternalistic, and patrons will be upset when Facebook likes don’t work? • Can we at least raise awareness, e.g. with tosdr.org plugin?
- 38. Example: smartphones • I DON’T EVEN KNOW, folks. • Smartphone owners do not control their phone’s privacy/security. Either Apple or their carrier (Android phones) does. • Phones leak data all over the place! • Location data particularly, but all “metadata” is of concern. • I don’t see an answer except better law. • Carriers are constrained by current legal framework to keep metadata indefinitely!
- 39. Bottom line: • Libraries and archives generally do privacy right. We certainly care about it! • A lot of online businesses are doing privacy very, very wrong. • Not to mention the feds! • And a lot of regular people are in no position to navigate the hazards. • So we have a serious problem on our hands! • And we owe it to civil society to continue to set a good example.