#critlib 2018-04-24: bias in web searching

Archive of the April 24 2018 #critlib chat on bias in web searching, moderated by @IamLibrarianSam and @metageeky. For more info about #critlib, including past and future chats, visit http://critlib.org.

  1. Looking forward to next Tuesday's #critlib chat! If you haven't read your copy of @safiyanoble's Algorithms of Oppression yet, maybe make it your weekend project? It's a quick, compelling, and necessary read.  https://twitter.com/IamLibrarianSam/status/985956589944365057 
  2. Looking forward to #critlib chat tomorrow (Tuesday) about bias in web searching, moderated by @IamLibrarianSam & @metageeky. Please join in even if you haven't read Algorithms of Oppression yet! 6pm Pacific/9pm Eastern  http://critlib.org/bias-in-web-searching/  #InfoLit #AIethics
  3. And in case you've read Algorithms of Oppression but want a quick look into issues of bias in search tools, I highly recommend Matt Reidsma's work on Algorithm Bias and Auditing search  https://matthew.reidsrow.com/  #CritLib #InfoLit #AIethics  https://twitter.com/violetbfox/status/988473698704416768 
  4. Bias is a topic I'm really passionate about, so I am very excited to co-moderate tomorrow's #critlib chat on bias in web searching! Be sure to checkout the questions and join the discussion!  https://twitter.com/violetbfox/status/988473698704416768 
  5. Less than 3 hours until tonight’s #critlib chat on bias in web searching, co-moderated by me & @metageeky! All are welcome to join, even if you haven’t read/finished the book! I’ve got my heavily marked up copy of #AlgorithmsofOppression ready to go! https://t.co/gdbfgPXvwC
    Less than 3 hours until tonight’s #critlib chat on bias in web searching, co-moderated by me & @metageeky! All are welcome to join, even if you haven’t read/finished the book! I’ve got my heavily marked up copy of #AlgorithmsofOppression ready to go! pic.twitter.com/gdbfgPXvwC
  6. @brantellsworth @metageeky Yes, Twitter! There's always a critical librarianship (#critlib) theme discussed. Go to  http://critlib.org  for more info. Anybody who is interested in critical theory and librarianship is welcome to join in on the discussion!
  7. Hi, MLS student here! Would it be alright if I joined in on the #critlib chat tonight, @IamLibrarianSam?
  8. @ashleighmaria Of course! Please do! Use #critlib when you reply to the questions we ask. Welcome to the Crit. Lib. community!
  9. So looking forward to the #critlib convo tonight. Who would have thought learning about search bias is more exciting than chilling in 70 degree weather!
  10. Introductions
  11. Hi #critlib! It’s time for our chat on bias in web searching, moderated by myself and @metageeky. Before getting into the questions, let's begin by introducing ourselves. I'm Samantha, an instruction librarian at a college in central Pennsylvania. Who else do we have tonight?
  12. Remember that #critlib has an Anonymous Relay so that anyone can participate. You can find it at  http://critlib.org/anon/ 
  13. Howdy #CRITLIB I am the Program Assistant for Learning and Outreach at Texas A&M University Libraries! pic.twitter.com/ZrykBh7SEw
  14. Hi! I'm Kate, the other moderator. I'm the Inclusion & Accessibility Librarian at Syracuse University in update NY. #critlib
  15. @IamLibrarianSam @metageeky I'm Ashleigh Florida-Young. I am a Masters of Library Science student in Middle Tennessee. #critlib
  16. I'm Sarah, grad student at UIUC - just listening in tonight #critlib
  17. Hello #critlib I'm Nicholas Schiller a former academic librarian (and still an academic) working in an archive on campus.
  18. Hi! I'm Symphony and a research librarian in the midwest #critlib
  19. Hi! I'm the journalism and govt info librarian at Columbia University #critlib
  20. Hello, I'm Jenny. I'm a librarian at Berkeley City College. #critlib
  21. Hi all, Charissa Powell, instruction librarian in Tennessee. Mostly lurking because I was at a conference all day! #critlib
  22. @sarahmkimes Listeners are more than welcome! Feel free to join in if you have anything to add. Welcome! #critlib
  23. Greetings! I'm library coordinator of @learningcommons at @TheLWTech. I'm especially excited to be learning with y'all tonight in anticipation of my presentation at @elunaorg next week. #critlib
  24. Hi #critlib - Hale here - academic reference and first year outreach librarian (for 6 whole weeks now!) in the Berkshires...
  25. If you're curious about tonight's questions or need a reference, they can be found here:  http://critlib.org/bias-in-web-searching/  #critlib
  26. Q1. What are some examples of problematic bias in web searching? What harm is done by bias in searching?
  27. Looks like we have a variety of perspectives here tonight! So let’s get started with our #critlib questions…
  28. Q1. What are some examples of problematic bias in web searching? What harm is done by bias in searching? #critlib
  29. A1: One of the earliest problematic searches on the web was "jew" giving anti-semitic hate sites as top results. #critlib
  30. A1. I'm mostly curious about how the act of using the tool doesn't encourage thinking about the tool. This is reinforced with the developers but also the culture around the "one stop shop" (Googlization). Cycles of bias inherently perpetuated. #critlib  https://twitter.com/IamLibrarianSam/status/988947709527756801 
  31. A1. #critlib I have seen @sarah_lemire talk with students about search bias by having them search 'genius'. How much do you have to scroll before you see a female or person of color? pic.twitter.com/7vBoQsweh4
  32. The argument was that those sites used "jew" instead of 'Jewish', hence their higher ranking. This led to an early googlebombing in which many people linked to a Judaism culture site with 'jew' as link text. #critlib
  33. A1 Predictive text for query inputs is the first thing that comes to mind. Esp those that rely on models based on previous search histories instead of progressive ones #critlib
  34. @LibraryMedlin @sarah_lemire I suspect that the results are just as bad for searches like "engineer" or "computer scientist" #critlib
  35. A1: My go-to examples are the SERP ranking of the stormfront-run martinlutherking [dot] org and the now corrected google image searches of Professional hairstyles for work vs. unprofessional hairstyles for work.  http://bit.ly/2r1EgHM  (SERP = search engine results page) #critlib
  36. @bembrarian I love that first sentence! Somehow we've become so passive in the ways we conduct searches and in how we are exposed to information. #critlib
  37. @LibraryMedlin @sarah_lemire This is a fantastic example. I hope we can compile a list of search experiments out of tonight's chat! #critlib
  38. There's also the problem that most programmers that set up the algorithms are people (most programmers are white men) with the own built-in biases, not necessarily intentionally racist, but it lacks a voice for POC. #critlib
  39. @midnorthwest Visualizing our selves as our entire search histories. Quite potentially terrifying, actually. #critlib
  40. Another good example of search bias is regard to various health topics. There's so much anti-vaccine/autism-fear out there that those results tend to rank high out of popularity. #critlib
  41. I currently teach high school English full time (librarianship part time). During my workplace writing unit, I attempted to explain the difference between business casual and business professional dress and all the images were of white people - when my students were not. #critlib
  42. Example: when you Google search "POC abbreviation," the first result isn't People of Color, but instead it's the last suggested result. #critlib
  43. A1: When I was trying to figure out how to have sex as a queer person it was REALLY hard to find results that were actually helpful and not just porn made for straight, cis, white guys... #critlib
  44. It's important to recognize too the difference between commercial web search like Google versus more constrained discovery systems like for catalogs, journals, etc. #critlib
  45. @nnschiller I use the hair example in my classes all the time to start a discussion about Google bias. It really get's students interested. #critlib
  46. @curlsinthelib Noble uses the professional hair styles in her book . . . with obvious bias preferring whiteness. #critlib
  47. Websites and Search Engine Optimization experts have more ability to game the system. #critlib
  48. This has evolved somewhat over the past 10 years, but not much, and not helped at all by biased web search environments... #critlib
  49. @metageeky Agreed! It's so difficult to keep up with how the search methodologies differ, but important to share (IMHO) that they do differ & that we should be aware of how they function. #critlib
  50. @haylmail This is a great point. Searching as a mode of education can be very, very personal and private, and often the information being sought isn't just intimate in nature but also is related to safety and health. #critlib
  51. @haylmail Oh gawd yes! Or everything gets slapped under SafeSearch or other protections because it's "icky" to the "norms". #critlib
  52. @metageeky Yeah, I think people forget that Google search is commercial or they think they can just ignore the ads and forget that there's more to it. #critlib
  53. @metageeky There's potential for cultural transformation (and equity) within both, or leveraging the users of each as a way to communicate societal issues. The developers have the opportunity to serve as models for change. But who's making change happen right now? #critlib
  54. Time to move on to the next question, but feel free to keep discussing Q1 as well. #critlib
  55. @jennyyaplib @metageeky My basic 101, out of the box, lesson that I repeat as a mantra to searchers: "Google is not magic." #critlib
  56. A1. As far as harm done goes--being passive users of information on the web means our worldview slowly but surely changes. Ultimately, our values and actions adapt accordingly, and it's feeding the cycle of bias. Research needs to be a mindful process. #critlib
  57. @jennyyaplib Even systems like DuckDuckGo, despite their promise of not tracking and the like, have money as a bottom line concern. #critlib
  58. @bembrarian @metageeky I think transparency is a great place to start for any internet search/discovery service but that is NOT the norm for most search platforms, esp. not the hyper commercialized. #critlib
  59. @metageeky @haylmail And what is really defined under SafeSearch, anyway? All part of the secret algorithms created to protect profit by corporations like Google.  https://support.google.com/websearch/answer/510?co=GENIE.Platform%3DDesktop&hl=en  #critlib
  60. @metageeky Oh no! I thought DuckDuckGo was an okay one! So what are our options? #critlib
  61. Q2. As creators of information, how can we be sure problematic biases don’t make their way into our own creations and work?
  62. Q2. As creators of information, how can we be sure problematic biases don’t make their way into our own creations and work? #critlib
  63. A2: I thank all the goddesses for colleagues/friends who proof-read my writing. They know what I mean, but won't let me get away with not saying what I mean. (all errors are mine) #critlib
  64. A2. Keep our own biases in check, and be thoughtful of the language we use when searching AND when creating out own content. Also, constantly be asking ourselves what perspectives are being left out of our results. #critlib
  65. @jennyyaplib @metageeky And ads can perpetuate cultural norms as well. Unless there's social movement to react to the system (like the case with Breitbart's retargeting ads), but how much power does it take to influence the ad network and its system?! #critlib
  66. Seriously though, check in with colleagues. Ask for input from co-workers. Don't work in a vaccuum #critlib
  67. A2: Be self aware first, and understand how that places you in the hierarchy of privilege - SELF REFLECTION first, and build empathy from there through education and conversation #critlib
  68. @metageeky Sadly, I don't think there's any way to 100% to avoid biases. In many cases, they're so ingrained to who we are that we don't know they are there. As much as we can, I think we must educate ourselves outside our comfort zone. Get opinions that make us uncomfortable. #critlib
  69. @bembrarian @metageeky @haylmail It's not a completely accurate answer, but in a lot of ways safesearch is keyword search w/o pagerank. There are blacklists as well, but basically... #critlib
  70. A2 Test test test. Doing things like Matt Reidsma, Safiya Noble, and others do. Test the topics you know about. #critlib
  71. A2) my first thought was “we can’t” - but I think it’s important to make sure lots of other people are reading my work, checking my, and that I need to be open to being checked. #critlib  https://twitter.com/metageeky/status/988951250128515072 
  72. A2: First, we have to be aware of our own biases. We need unconscious bias training just like anyone else - no matter how much of a minority we may be ourselves or how interested in equity/inclusion we are. #critlib
  73. A2b. In truly being transparency, and thinking critically, we can also design created materials to explore biases directly, and acknowledge where they may exist. Opens the opportunity for conversation, dialog.  http://learningoutcomesassessment.org/TransparencyFramework.htm  #critlib
  74. A2: Awareness of the politics of citation (invoking Sara Ahmed) in our own research practices is critical. Who is cited? Who and what are being invisibilized and under/mis represented in our information production and research? #critlib
  75. #critlib hi, Todd here. Academic librarian at an urban community college. A1: This is so important. Most of my students don't realize that their Google results don't look like mine. Bias is built into every search based on what you've clicked on in the past.
  76. @haylmail I agree! Self-reflection is necessary to have these transformative kinds of learning experiences. If we don't work on ourselves, how can we be expected to help our patrons and communities? #critlib
  77. Let me clarify. DuckDuckGo is largely donation driven but is still a business. Their emphasis is on privacy and not tracking users. That's good, but that says nothing about potential bias in their search results. #critlib  https://twitter.com/jennyyaplib/status/988951871430656000 
  78. @ashleighmaria @metageeky Well said! Bias (in the most technical sense) is unavoidable, but that doesn't mean we should stop aspiring to do better. #critlib
  79. @jennyyaplib @metageeky Though I agree they (and no one) is perfect, I have noticed the leadership of @DuckDuckGo being very open to discussion, dialog, and critical feedback in a way mega search corporations like Google and Bing are not. #critlib
  80. Hello! Cataloger in Minnesota. I am joining late because I got stuck on the phone with my mobile carrier. My problem was not resolved, grrr, sigh. #critlib
  81. @CharissaAPowell I've caught a spelling error of mine already, too. Just roll with it. :) #critlib
  82. A2 Also, report back to the vendors. Challenge them. Make them have to develop solutions that are more hardcoded fixes for some problems. #critlib
  83. A2 I think it's important to begin teaching undergrads how scholarly publishing works; we talk a good game about peer review & bias, but how often are we imparting to students that the system in place is seriously problematic? #critlib
  84. A2c. I'm really interested in this idea that we all assume everything has to be indexed by search engines. That's technically not the case. We should be aware that search indexing is a choice as well, and choosing to index or not both have implications. #critlib
  85. A2. I also think it's important to note that bias is not inherently bad. It's part of being human. It becomes problematic when we are ignorant to it. #critlib
  86. @metageeky @bembrarian @haylmail I'm speculating, so I may be talking out my elbow here, but the links-as-votes mechanic of PageRank, doesn't seem to apply to content filtering. Maybe what I'm speculating is that what PageRank tells us is less useful in that context? #critlib
  87. This. We can build off what search results already exist. De-silo the search experience. Make content within search results community-driven. #critlib  https://twitter.com/metageeky/status/988952417877352448 
  88. @marijel_melo Agree! This is something I'm working on a lot as a writer and being super aware of who I am citing and who I'm leaving out has led me to the research places I really want to be at - don't take the easy way out! #critlib
  89. A2: Speaking of bias, in determining ARWU rankings, freaking 20% of the assessment criteria is based on WHAT journal an article appears in! #critlib
  90. A2: Adapt and champion platforms that are designed and supported by minorities in the tech industry #critlib
  91. @IamLibrarianSam Bias also inevitably leans towards providing depth & breadth for the dominant culture, though, so it's not like it's "innocent" even if we're aware of it. #critlib
  92. @librarian_lali Completely! Even though peer review is by experts in your field, under researched topics tend to get reviewed by semi-experts with lots of misconceptions. #critlib
  93. @librarian_lali It is my DREAM to teach an entire semester of information ethics. There's so much to cover, and it overlaps every discipline. #critlib
  94. @bembrarian But to be fair, you have to address the bias in the community too. Tyranny of the majority is not just a band name. #critlib
  95. Q1. I think it's important that the question says *problematic* bias. Because there's no such thing as neutrality, no absence of a perspective. It's necessary to be conscious of what your perspective is, and be thoughtful/intentional about it #critlib
  96. @marijel_melo I've learned to listen closely when people refer to the politics of citation. Fascinating conversation follows. Thank you. #critlib
  97. @metageeky Going back to ExL, a quick search for "diversity" on their website yields . . . what kind of documents? 😕I'm very much reminded of the movement from Uber to Lyft last year as a result of similar lack of initiative by the former to support equity #critlib
  98. Bias can be an important way of keeping oneself safe, but also can be over activated by stress, anxiety and just living in SOCIETY. Try to pause and question gut reactions... #critlib  https://twitter.com/IamLibrarianSam/status/988953872310919168 
  99. Ppl believe they *have* to accept platforms based on racist, sexist, homphobic algorthyms bc they have no choice. Let's show our users they have options #critlib
  100. Ironically, my Info Policy course's topic this week is ethics and the prerecorded lectures focused painfully on the trolley problem and the opinions of two privileged white MLIS students. #critlib
  101. @jennyyaplib @IamLibrarianSam @safiyanoble I recommend reading Dr. @safiyanoble 's work, and also looking at intersections between scholarly publishing and IL -- @StephDK and @mhensle1 have written & presented on this very topic (and included convos about the ACRL Framework for IL in Higher Ed. #critlib
  102. @librarian_lali Analogy: Who or what shows up at the top when you search for them in a search engine, in social media, in the database, Yelp, OKCupid, etc? And why? Authority in search rules the landscape. #critlib
  103. @metageeky Agreed. Being aware that it does exist or how to identify/think critically about it. Not inherent in most of the systems being referenced today, sadly. #librarylife #critlib
  104. Q1. My current job doesn't involve this sort of thing, but I wonder if a good exercise for students would be to search a bunch of different synonyms/ways of expressing something, then reflect on differing results. #critlib
  105. @midnorthwest Options, yes, and also options for reading search engines as they (hopefully) do any other document: critically. #critlib
  106. @bembrarian @librarian_lali I think that it's key to focus on how Web Scale search does authority differently than what came before. Using what people click on is still authority, but it's less conscious authority than an editorial board.#critlib
  107. Q3. How can we thoughtfully perform bias analyses on our own systems?
  108. As always, continue to discuss Q2, but here is the next: Q3. How can we thoughtfully perform bias analyses on our own systems? #critlib
  109. A3: First. Listen when someone reports a complaint. Seriously listen. Try it out. Don't dismiss it as not that serious. Ask relevant people for their own opinions if you're not sure. #critlib
  110. @curlsinthelib @SarahEHare @charbooth Also, look at the work of COAR (Confederation of Open Access Repositories) - they are trying to address open access and sustainability. #critlib
  111. @haylmail @midnorthwest Options and yet . . . the options continually seem limited. Baidu might be worth thinking about as well and the potential for search monopolization. #critlib
  112. And relevant is key here. There is a lot of heterogeneity in underrepresented groups. #critlib
  113. @aboutness I like this as a critical application. I have done this w. LC subject headings at the reference desk, but it is a good idea for ALL search environments. #critlib
  114. A3 #critlib I've been thinking about this since the ?s were posted. I'm phrasing it as a question: can (should?) libraries pay for sensitivity readers in their discovery tools in the same way that novelists pay sensitivity readers of their books?
  115. A black person might not be as sensitive to slurs used against Asian people. A wheelchair user might be indifferent to mocking of mental health. #critlib
  116. I experimented with this while working on the SAC Working Group on the LCSH “Illegal aliens” report, and the results for various terms (undocumented immigrants, foreign nationals, irregular migrants, unauthorized aliens, etc.) are dramatically different. #critlib
  117. @haylmail @aboutness Yes. And as painful as it may be, we might have to keep around "bad" subject headings to aid legacy searches or biased patrons. #critlib
  118. A3 #critlib I know it's a separate problem and I should compartmentalize, but it's hard to separate bias in our collections and our controlled vocabularies from bias in our search tools.
  119. A3: Take to the community you serve. Ask them to provide critical feedback verbally or in an optional survey. Make the community feel valued by valuing the community. Reach outside the community for a different viewpoint. #critlib
  120. @metageeky Also important to not make relevant minorities do all the work of educating bias away. #critlib
  121. @nnschiller @librarian_lali SEO folks are starting to assume rank is also influenced by the volume of users accessing the site directly (by entering their URL into the search bar), based on Chrome data and Analytics. The search experience begins before and beyond search. Bias is very total. #critlib
  122. This is a great question, but one aspect I see as different is that being able to test search results is a very different task that reading a novel. #critlib  https://twitter.com/nnschiller/status/988956719765073920 
  123. @nnschiller This is right where my brain goes, too. I can't separate the two. When I teach students about controlled vocab, I wish I had more time to really get into Google vs. library search tools, and how both have faults. #critlib
  124. @nnschiller Ugh, I'm struggling with the idea of paying sensitivity readers when knowing the corporations probably aren't going to pay them. #critlib
  125. @metageeky @haylmail As "See" references, you mean? Even that needs to be done with extreme caution. Some terms are so harmful they need to just be gone. #critlib
  126. @metageeky @aboutness Hence the responsibility to be ready to explain violent, harmful systems and the bias that created them. Information professionals need to own this and be ready to dig into the uncomfortable-ness of these spaces. #critlib
  127. @nnschiller Oh, no, I can totally see why you would though, though. We keep waiting for our tools to get better so they can make better use of our vocabs, but... our vocabs aren't always great? All are symptomatic of the same bias. #critlib
  128. Also, it's important to not put the removal of bias solely on the backs of underrepresented groups. And then there's the tendency to use that as a not my problem excuse: "They said it was fine!" #critlib
  129. @metageeky @haylmail @aboutness *might* being key term here. Some might not be afraid of addressing biased patrons with a variety of teaching and learning moments. But I respect those who are afraid. #selfcare #critlib
  130. @midnorthwest What a meta concept. I didn't think about this, but yes--having a common understanding of how we define bias would be a good place to start. #critlib
  131. @metageeky @aboutness Also it def. feels like the successful research of biased patrons is OFTEN privileged over the harm done to minority researchers. #critlib
  132. @midnorthwest I like the idea but it's sort of like defining pornography. And if we go too far in making if very academic, the definition may not be of practical use. #critlib
  133. @IamLibrarianSam @nnschiller But at least library search tools are the evil we know, whereas Google is an algorithmic black box. (As discovery layers move that way, we give up even *that* advantage, of clarity & transparency.) #critlib
  134. Q4. What can we do to help our communities understand the biases that exist in web searching?
  135. We're beginning to reach the end of the hour, so here's the last question. Q4. What can we do to help our communities understand the biases that exist in web searching? #critlib
  136. A4. For one thing, don't be like that an IEEE idiot. #critlib
  137. @IamLibrarianSam @nnschiller Scale becomes something I can't stop thinking about. The scale of search engines (in most cases) overshadows any particular library database/collection. And as such specific tools with specific inherent biases are used to algorithmically maintain search engines. #critlib
  138. @violetbfox @IamLibrarianSam I think that distinction is already lost in the current generation of discovery layers. Google, at least, has some error-correction built in, where I don't think discovery tools learn from mistakes. #critlib #cynic
  139. clarity that we have through our own professional privilege though - for the average user library search tools are not much more transparent - this comes through relevant instruction. #critlib  https://twitter.com/violetbfox/status/988958671131631616 
  140. @IamLibrarianSam A3. If possible, dismantle, break, pull apart the systems to see its insides and how it works. Whether it's a coding language, algorithm, or 3D printer its parts reveal much of the socio-economic, gender, class, race biases that comprise the system. It's all baked in. #critlib
  141. A4. Another thing I think we should do is push back on the tendency to make things "work like Google" #critlib
  142. A4: Supporting those who have curated examples like @safiyanoble and curating examples ourselves when we find them to help us break down the idea that data is neutral #critlib
  143. A4: At the reference desk, I think we need to constantly talk about how the tools work as we use them. Make it a kind of carnival/salesperson patter, talk about what's going on. #critlib
  144. Q4. I spend a lot of time thinking about this question. I'm an instruction librarian, so my "community" is primarily college students. I've found the best way to help them understand bias in searching is by providing examples. #critlib
  145. One example I use a lot is searching "Iran Hostage Crisis." I ask students why searching this wouldn't include all perspectives. (Answer: Iran calls it "Conquest of the American Spy Den.") #critlib
  146. @metageeky This is key. Thank you for sharing this tonight. There is a role for #whiteness in a lot of this work. I love the work of Robin DiAngelo, where I've started to think about my own position. #critlib
  147. A4. More people seem aware of data privacy concerns in social media than of search engine bias (and how profit motive impacts Google etc. results). They are not totally unrelated, maybe there's a way to help people see/understand the connection? #critlib
  148. A4 And to at least give some credence to the IEEE guy, explain how much tracking and personalization Google does to your results. #critlib
  149. A4: give up your power by admitting that YOU are biased b/c you are human. And point on that HUMANS make search tools and the algorithms that run them at every possible moment. #critlib
  150. Even private browsing in Chrome doesn't remove you from Google's tracking. They know that browser is coming from the same IP address as before. #critlib
  151. @metageeky Totally agree. But I think institutions can come up with a definition that reflects their specific orientation to bias while still following some broad, theoretical oh no, dare I say it, framework? #critlib
  152. A4: The one thing I've noticed really matters these days is the language you use. It's so easy to "trigger" someone with buzzwords and suddenly they completely shut down and won't listen to what you say when you bring up social problems in the community. #critlib
  153. @haylmail Yes, you're right, but I want to give libraries credit where credit is due for having processes which are at the least "knowable," if not always "known." 🙂 #critlib
  154. A4b. But really, have conversations, focus sessions, workshops, instructions, and all other methods of communication that bring this topic up regularly! This is information literacy! This is critical thinking! It should be normal--NORMAL--to talk of systemic info biases. #critlib
  155. A4a. "Dear [name of community], we regret to inform you that the search engines you use and the content within are biased. It's all biased. Please chat with your librarian for more information." #critlib  https://twitter.com/metageeky/status/988958798500032514 
  156. Yes. In teaching about privilege, I believe there's a fine line in making people uncomfortable but not so uncomfortable that they put up defensive shields against hearing more. #critlib
  157. @metageeky Yep, we're logged into the browser in most cases. Or the browser session/tab. And then all of the devices. And all of the devices' connected apps. #critlib
  158. @aboutness Dr. Noble has talked about the commerical bias of SERPs and search engine rankings. (Vogue and Glamour rank above feminist magazines because the latter feature fewer or lower value AdWords.) Maybe? this is the source  http://bit.ly/1953BW3  #critlib
  159. So #critlib, we are passed the hour. Thank for attending tonight and thank @IamLibrarianSam for proposing this topic.
  160. @metageeky Addressing somebody else's bias takes tact. We need to give them time to be uncomfortable with it, and give them a chance to eventually come around to it. Likewise, we need to lean into our discomfort when somebody disrupts our biases. #critlib
  161. It's been almost two years since this came out, and still no response/action from LC. If you are interested in a very deep examination of a single problematic/biased subject heading, please do check it out! #critlib #critcat  https://twitter.com/aboutness/status/756282780045103104 
  162. #critlib hour always goes by so quickly! Bias is a topic I've always been so interested in, so thanks for all of the great perspectives tonight, everybody. If you have any announcements, go for it!
  163. @metageeky @IamLibrarianSam Thank to you both for hosting and everyone else for participating! Fantastic thoughts from all. #critlib
  164. @bembrarian Yes - this. Work it into every session! Be a broken record if you have to. #critlib
  165. @aboutness Makes you wonder if there's something with more teeth that could be done to push this through. #critlib
  166. The best part of #critlib chats--the entire hour. The worst part--after the hour is up and I go back to re-read everything, only to realize how many spelling errors I made. #fail