Libraries and Learning
Evidence and Authority in the Age of Algorithms
(Presented at “Teaching Writing in a Post-Truth Era,” University of Notre Dame, August 20, 2019)
I come to the issue of teaching writing in the post-truth era from a somewhat different perspective than our previous speakers. I’m a librarian who has long been interested in the ways students get ideas, interact with other’s ideas, and how their experiences as writers in college shape their identity as people with agency and a grasp of how knowledge is made and negotiated by people – people like them. I’m taken with the parallels between writing instruction and what librarians do. Your writing program has as a goal ethical and moral use of words and evidence.
Making an argument is an ethical activity, one that helps students develop intellectual and moral virtues.
It’s about learning how words work and how to use those words ethically. This is also what information literacy is about. Learning is the primary purpose for librarians’ work with undergraduates. My library’s definition of information literacy is similarly ambitious – not just how to find and use information in the library and online, but more deeply to understand where information comes from, how it’s connected to social processes, and how they can participate in those processes with a clear sense of right and wrong.
We sometimes get distracted from these larger goals, though, because what’s on students’ minds is super practical: how to complete an assignment that’s breathing down their neck. How to find those five scholarly sources required. It’s a legitimate concern, especially in this age of anxiety when curiosity and exploration seem like a luxury. When you’re trying to figure out where you are in this new academic setting, just surviving may take up all your energy. When we talk about “student success” we’re not necessarily talking about lifelong learning and the purpose of liberal arts, we’re talking about the ability to perform well as a student, to write academic prose, format footnotes, and draw from scholarly sources without getting in trouble for plagiarism. That tends to push loftier goals to the side.
Wrestling with big ideas and with other’s thoughts using methodical forms of argument can be an incredibly valuable learning experience, but it’s the beliefs that underlie academic inquiry that are really important. Searching for truth ethically, with an open mind, without cherry-picking evidence to “prove” a preconceived answer, learning to live with complexity and ambiguity and feel some responsibility for doing that work – we need that kind of thinking badly. But I’m not sure that when students learn to write academic prose, citing academic sources the way scholars do in their publications, that the values implicit in that kind of writing are clear to them. And while this is nothing new, it seems especially pressing in this post-truth era, when words like “argument,” “proof,” “evidence,” and “truth” are more slippery than ever and so often abused.
I have always wondered what parts of students’ academic experiences matter after college. How does my work with students provide lasting value? Are they simply learning to use my library and our databases, are they becoming somewhat fluent in the use of academic phrases and styles of argument so they can do well in courses? Or do the values underpinning the specialized and complex language of researched writing actually reveal themselves to students as they do this work? I hope so. As they learn to think like historians, like scientists, like artists by engaging in the practices and methods of those fields they may learn along the way what makes for honest, rigorous inquiry – and why those values matter. By writing and conducting inquiry in a disciplinary framework they often gain the understanding that knowledge is a thing we make together, ideally learning that they have the agency to be a participant in making the world as we know it, and that they think about it as ethical activity that engages intellectual and moral virtues.
But there’s troubling evidence that students compartmentalize the kinds of evidence wrangling and argument-building they do for academic tasks and what they do in their personal lives. I suspect the complexity of mastering academic writing can often obscures the basic moves of academic practice.
For example, mastering APA or Chicago format can take up a lot of head space – so much so, that many undergraduates fail to see the utility of citations or know how to use their coded information to find information. Reading citations as a map of ideas is second nature to us, but they’ve grown up in a world of search, where the terms you type in will provide answers based on what you appear to want because the platforms they use are built with the assumption you are consumers shopping for things – and ideas are simply another kind of thing.
It’s difficult for students composing citations to see from that work that knowledge is a conversation, an ongoing collective process involving people wresting with ideas together over time. My sense is that academics have so internalized the social processes and values of research that they forget the social nature of knowledge is not obvious to students. When they hear “include at least five scholarly sources” they don’t hear “because” – instead, it’s a shopping list, with possibly the name of the database-store they should go to (or the shops they should avoid). When they create a list of references, it isn’t a map of ideas and how they connect, it’s a list of ingredients required by law, fine print nobody reads but the instructor, looking for technical errors.
What does this have to do with post-truth? That tacit knowledge about how we try to arrive at the truth is incredibly important and is easily lost in the current information environment. We need to catch up to what’s happening with information and bring that understanding into the classroom, into the ways students learn about inquiry and argument. This year I’m fortunate to be working on study with Project Information Literacy, an independent non-profit research institute that uses social science methods to explore student experiences and beliefs about information. This is the largest body of research on how college students interact with information. In a report we plan to release in December, we’ll be summing up ten years of findings and discuss where we go from here.
Some of the things we’ve learned from a decade of PIL research so far:
- Students find getting started especially challenging and time-consuming. This hasn’t changed over the years. Notice, one of the reasons misinformation spreads so easily is lack of context.
- Most students use a small compass to navigate the plethora of information resources, using the same path – efficiency, utility, predictability. Google has made search incredibly convenient. So easy, you can ask your phone a question and get an answer in less than a second. If it takes time, students are inclined to think they’re just doing it wrong, that they just didn’t find the right search terms to find the perfect source.
- While a large majority of recent graduates surveyed believed they had transferred information skills from college to their lives now for searching and interpreting and applying search results far fewer agreed that college had helped them develop the ability to formulate questions of their own. Employers said recent grads had trouble seeing patterns and connections, and were reluctant to take a “deep dive” into a variety of information sources; the efficiency they honed in college to meet deadlines led them to skim the surface.
- The most recent study about how students engage with news found students learn about news in their classes and from their professors as well as from fellow students. They care about current events and believe news is necessary for democracy and being informed is a civic duty, but have trouble navigating the deluge of news and knowing what to trust. Half of students said it was hard to know what was true and a third said they didn’t trust any news sources.
We’re embarking on a new qualitative study. We’ll be summing up what we’ve already learned from ten years of research to look for gaps in our current understanding, we’ll be asking students and faculty what they know and think about the ways information flows are being shaped by “black box” algorithmic systems, and will be making some recommendations about what we might want to do to help students understand how information works in our and how they can participate in sharing and creating information responsibly as citizens in a changing and challenged world.
We want to get a handle on what students know and think about the fact that so much of the information we encounter comes through platforms that were designed for selling things – and not just consumer goods, but selling a political candidate, a political philosophy, or a message of hate. These platforms have honed technologies of marketing manipulation that are effective because they are able to gather and process massive amounts of personal data, and they can conduct constant experimentation to see which messages have the most power for micro-targeted audiences. These are platforms our students use and know lots about – but we so rarely connect those experiences to what happens in the classroom. When we say “Age of Algorithms,” we’re not just taking about computer code, we mean an interlocking set of technical and social developments –
The systems we interact with most often throughout our day now that we carry computers in our pockets are built to keep us engaged, to stay on the platform clicking and sharing. They need that engagement not just to serve ads, but to gather our data to personalize what we see. They want to know enough about us to predict our behavior (whether for shopping or voting), and they are able to nudge us toward the behavior they want without our being aware of it. Shoshana Zuboff calls this the Age of Surveillance Capitalism and is really alarmed about the implications for democracy and for human freedom and dignity in a world where these companies have so much power and wealth based on monetizing our lives, when we know so little about how we are being used.
In addition to platforms that we see and use regularly, there are some behind-the-scenes technological developments that are part of the Age of Algorithms. It’s possible now to rapidly and at vast scale collect and process fine-grained relational data in real time. That data, though, used to train artificial intelligence systems is often incomplete or biased, so they learn and actually amplify bias. This has an impact on who can rent property, get a job, or a long prison sentence. It also affects search engines, as Safiya Noble has pointed out in her book Algorithms of Oppression.
The scale of it all is pretty breathtaking. There’s just so much stuff being produced and shared, and its provenance is so hard to figure out, it’s really hard to ask those basic rhetorical questions: who is speaking? Who is the audience? What is the speaker’s purpose?
It’s not just technology, either. The giant corporations that dominate the list of largest global companies by market capitalization seem unable to anticipate or respond to unintended consequences, behaving according to deep roots in Silicon Valley culture, – indifference to or ignorance of perspectives different than those of affluent white males, a false belief in meritocracy, a global reach coupled with a lack of cultural understanding, and fundamentalist faith in individualism and absolute free speech. They don’t turn to the scholarly literature about technology and society because it doesn’t occur to them that scholarship can have value. A technology writer recently complained about how annoying it is for scholars pointing out the publications that address questions technologists pose. Scholarship is irrelevant because ideas are not valuable to “doers,” and besides who has all the money and power?
Bringing this closer to the classroom, the disaggregation of published information and its redistribution through search and social media platforms makes evaluation of what used to be distinct sources, like articles published in a particular journal or stories in a local newspaper, all the more difficult. This leads to an individualized presentation of information that sorts results based on inferences drawn from personal data trails. We don’t all see the same information when we search and it’s not obvious where it came from. Decades of media consolidation, deregulation, and economic trends have been exacerbated by the rise of social media platforms that are designed for persuasion in ways that has contributed to engineered distrust of traditional knowledge traditions and the destabilization of political and social institutions globally.
But there are precursors to the technologies and contemporary culture of the Age of Algorithms that have been enormously influential in creating our post-truth moment. Our news environment has been bifurcating since lobbying ended the Fairness Doctrine and aided the rise of right-wing talk radio and Fox news network, which is a self-referential system that places less value on fact-checking than on adhering to and reinforcing a world view shared by its white conservative audience. Network Propaganda is a study of how supporters of Clinton and Trump shared news between 2015 and 2017. By creating maps of sharing patterns on social media and on the web, you can see the insularity of the right-wing news sphere. Information from the far-right fringes was picked up and amplified by Fox, with its large audience. Information on the far-left fringes failed to gain as much traction because its news consumers also valued mainstream news sources and traditional journalistic values and traditions – fact-checking, confirmation, issuing corrections when warranted. The study’s authors concluded that what mattered most wasn’t Russian interference or the impact of algorithms and filter bubbles but rather what they call a “propaganda feedback loop.” What matters to the right-wing audience is consistency with their overall belief system. Whatever information far-left-wing publications presented was moderated for its audience by their reading of the mainstream press. Left-wing audiences rejected ideas that were contradicted by factual findings and traditions of traditional journalism. As New Gingrich explained to a rather baffled interviewer, it didn’t matter if it was a fact that crime was down. What mattered was whether people felt crime was down. Feelings are factual; facts are merely theoretical. This rejection of facts is reinforced by declaring the press that isn’t part of the right-wing feedback loop untrustworthy by definition.
So what we’re faced with isn’t a technology problem that can be fixed with technology. It’s a social problem that has been exacerbated by a technological architecture of persuasion. If truth is not something you value, if all you care about is selling an idea, you’ll have full command of the power of that architecture as we have seen, over and over, as the same inciting words and phrases circulate between 8chan, talk radio, Fox, the president’s Twitter feed, and the statements of mass murderers. Authority is redefined as adherence to a belief system rather than to methods or codes of ethics.
What makes it even more complicated is that evidence itself is embraced as material to be woven into coherent narratives – conspiracy theories in which individuals feel empowered to scour the web for material to fit into a narrative. “Self-investigation” is a way of dethroning traditional authorities while building a case for a flat earth theory or against vaccines, only calling on authority if it confirms the theory (which is why it’s so valuable for fringe elements to win coverage by mainstream media).
All of this makes it hard to know what to do. I’d like to leave time to share ideas about that, but here are some starting points that
- Connect and contextualize current events – look for ways to bring news into your courses and connect those events with disciplinary knowledge and knowledge-making methods. So much of our susceptibility to false narratives is rooted in lack of context. Connecting disciplinary ways of knowing and habits of thought to reading current events can also demonstrate the applicability of academic ethical inquiry to life beyond college.
- Model ethical information practices beyond academics – students in the news study assumed the driving motivation of journalists is to sell ads. They had no sense of the traditional divide between the newsroom and the business side of news organizations. They didn’t know journalists have a code of ethics. Some of that is youthful cynicism, but some is simple ignorance about how different kinds of information are created. So explore how peer review works in practice, how news is reported, how opinion and news have different editors and different functions, how it’s just as important in everyday life as in the practice of history or science to handle information with care and integrity.
- Build trust in info institutions and practices – we’re pretty good at what Peter Elbow called “the doubting game,” but we also need to help students believe in the value of truth-seeking institutions that have ethical values and tested methods for avoiding prejudice and falsehoods. By all means, point out failure points – the retracted science article, the news story that turns out to be horribly wrong – but also take some time to point out the social processes and values underlying good scholarship, good reporting, good and honest writing of all genres.
- Practice information encountering skills, not just search skills – we learned from Project Information Literacy’s news study that students feel continually bombarded with news, or as one student put it “News just throws itself at you. I don’t try to follow the news at all but it still throws its ugly self into my face on the daily.” When it came to school assignments, students sought sources, but were far more casual about evaluating sources they encountered. I think it could be helpful to discuss how you encounter sources in your own field and how to develop a curatorial sense for shaping the news feeds you use to keep up with both disciplinary knowledge and news given so much of what we know is thrown at us, not sought.
- Practice practical evaluation skills frequently – The heuristics we have typically used to help students evaluate sources don’t actually work in an information abundant environment. Many students get the message that scholarly sources are always safe, JSTOR is always a sure bet. Not everything we need to know is found in peer-reviewed research, and a lot of peer-reviewed research is not only inaccessible to non-specialists but not high quality. Using “peer-reviewed scholarship” as a binary sorting machine is tempting but not helpful in the long run. Neither are checklists that ask students to answer questions about a single source out of context. Is the author an authority? Is the publication reputable? How could students possibly know? Sam Wineburg of the Stanford History Education Group led an extensive study that found students were pretty bad at evaluating digital sources. A second study, though much smaller, was even more interesting. He asked historians to evaluate digital sources and they were bad at it, too – close reading and primary source examination don’t really work for the web. Fact-checkers were much better at it. Drawing on those insights and the way digital works, Mike Caulfield has argued we need “web literacy” – easy heuristics for understanding information we encounter online that can help us make judgements quickly and without going down rabbit holes. His latest iteration of four moves is available in five adaptable 30-minute lessons especially useful for an introductory writing or information literacy course, but I would suggest using them in many settings, repeatedly, to give students critical habits that are easily transferable across domains.
Those are my first thoughts about what we might do, and we can carry on the conversation. The main point I wanted to make is that our information environment has changed and is now optimized for spreading propaganda, disinformation, and creating a sense of helplessness, a sense that nothing can be trusted. We have to offer our students hope and build bridges between a clearer understanding of academic values and how we live in the world so that students are ethically and intellectually equipped to work to make it a better place.
image sources
“Human computer” (1940s?) via NASA
Two women programming the ENIAC computer (1946) via Wikimedia Commons
Two men working on a Census Bureau computer (1950s) via Wikimedia Commons