Could an information architecture foster a place to dwell and reflect?
There is no doubt that we continue to be in a state of misinformation and disinformation, globally. In Five takeaways from UK’s AI safety summit at Bletchley Park, Dan Milmo and Kiran Stacey note, “…if there is a consensus…then it focuses on the immediate fear of a disinformation glut.” Certainly governments acknowledge the dangers of disinformation and the need to create policies to address it. The voting population needs to be aware of this and educate themselves on media literacy. The naive answer to mis- and disinformation is to provide true information, to validate statements as factual, to promote content that is true. We may not know how to do this given current incentive structures (for example, given algorithms deployed that results in conflicting priorities). But this has been the clear goal for many hard-working individuals on the front lines of content generation and social media.
Currently, we strive to provide a way to verify the truth value of some content. We promote urgent content. We provide useful—or usable—content. But is verified content, or the dispensing of metadata about urgency, usability, and other qualifiers, a primary concern to the media consumer? Content consumers may be engaging in mis- or disinformation, not in order to validate some rational position or even irrational ideology, but for other motivations. Here, I’d like to expand the notions of civics-building by looking at a more expansive view of information that may engage the content consumer.
Instead of focusing on news or statements or data points, and determining whether they can be validated, can civic engagement arise, or improve in quality, when the consumer is confronted with content categorized as different types? Calling out different types of content might empower the consumer to engage in civic dialog on a variety of levels, using more than just rational discernment. Categorizing the type of information for a user in a non-prescriptive manner, the consumer has the freedom to engage different cognitive functions and judgment processes, each tailored to the conventions, purposes, and truth-claims of the content type in question. My objective, here, is to identify and sketch out a matrix of qualifying content, not based on a truth value, but to nurture public engagement that is critical, self-reflective, and mature, resisting the forces of domination and fostering rational-critical debate.
To Kill A Mockingbird
A novel exemplifying civic integrity is To Kill a Mockingbird. When students are required to read this novel, they fall in love with each character and in turn primal feelings of fairness are triggered. The carriage of justice, while imperfect, is celebrated. They are introduced to universal truths declared by democratic founders, such as the notion of the singular value of each person and self-governance. So it’s easy to say that a novel can convey and inspire truths and ideologies without asserting any facts. This is one of the reasons why the recent book bannings are so disheartening to so many of us. Those who have read banned books recoil when we hear that advocates of taking these books away from young people have actually not bothered to read the books. Books, whether fiction or fact, memoirs or hypothetical worlds, can convey foundational truths. These truths serve up the basic moral questions on which an understanding of liberal democracy can be developed. These books can inspire civic engagement. Well-meaning crusaders against disinformation cannot imagine wading into the murky waters of fiction. But those who want to eliminate civic awareness, and democracy itself, are definitely pulling books off the shelves. It is just at times such as these when fictions can contextualize the present moment and set the stage for possible futures.
To Kill a Mockingbird is definitely part of the canon of secular humanism and liberal democracy. It is critiqued from the left to the right, reducing the sum of all of its good qualities by its countless faults. But one would be hard pressed to categorize this book as propaganda. There is a holistic crafting of personalities, time, place, connections, and relations that develop a place where the reader can dwell and reflect.
It’s no easy task to create a novel that has the proper mechanisms to trigger primal notions of fairness, respect for life, and a call to civil engagement that avoids propaganda. That sort of expertise is actually what makes good fiction good.
Orwell that ends well
Let me take a step back, though, and ask: could categorizing content types be counterproductive? There are plenty of examples of information (dis)organization deleteriously constraining civic dialog. And maybe even more interesting to ask: is there something inherent in at least some built-in content types that would allow them to function against manipulation, or even repression, that could arise by the system in which it is embedded? By building a robust structure that allows for multiple types to coexist, might we be able to construct a kind of constitutional durability, and even promote civic agency, within the constraints within the system?
There is hardly a more sensational example of implementing social constraint than the Chinese Social Credit System (SCS). Ordinary People as Moral Heroes and Foes: Digital Role Model Narratives Propagate Social Norms in China’s Social Credit System investigates the shaping of social behavior that upholds individual citizens as exemplars of sincerity in economic activity and punishes those who transgress. These narratives often align with the Confucian virtue of “chengxin” (诚信), sincerity or integrity, which is pivotal in Chinese ethical thought. But focus is on economic transgressions. This reflects a shift from a broad ethical narrative to a more targeted economic behavior-focused narrative. With materialist philosophy, economic actions and outcomes are given moral weight. This resonates with Marxist theories where economic bases are considered fundamental to societal structures and class relations. The publication and promotion of SCS is part of the plan to ensure widespread understanding and acceptance of the SCS norms.
The fear we all have in this discussion is that the use of SCS for state surveillance and control mirrors the panopticon, the constant surveillance that could induce self-regulation in behavior and limiting civic agency. And the labeling of individuals as “Lao Lai” (moral foes) without full justification may lead to mischaracterization and discrimination. So even if the ideal, future ideology is something that we might agree upon, the consequences of implementing an SCS could be personally devastating.
In the article Orwell that Ends Well? Social Credit as Regulation for the Algorithmic Age, Kevin Werbach puts forth an idea that China’s SCS could be separated from ideology and used for liberal democracy. Werbach notes, “Whereas in China the danger is political repression, in America it might be exacerbating inequality.” Because, “reducing SCS to Big Brother or Brave New World, however, differentiates it too starkly from the profit maximizing surveillance ecologies of the West.”
Social credit is, in the end, a manifestation of social capital. Unlike financial capital, social capital is a form of wealth that cannot be established by fiat. It must emerge from relationships and experience and communities. China’s SCS seeks to aggregate and analyze countless weak signals to strengthen patterns of connections. Its immediate objective is compliance, and within the Chinese context, compliance may involve injustice that liberal democracies appropriately protest. The ultimate goal of a social credit regime, however, is to promote a flourishing society. There is no reason that vision cannot be transposed to contexts in which a flourishing society equates more strongly with flourishing individuals.
The credit of SCS is applied against an individual’s account. But what if we applied some metrics against content and not the individual? What if we could quantify the civic agencies of fictional characters? Who then is harmed? But, even more to the point, by quantifying content, could we call out a particular piece of content as a particular type, so as to trigger the appropriate awareness in the content consumer? Knowing one’s rhetorical realm signals different faculties.
Most of the Western commentary on SCS is focused on the infringement of the rights and privacy of individuals. The enforcement mechanism produces “red” and “black” lists in an effort to orchestrate laws and norms and to mete out rewards and punishments. In China, science fiction is arising in popularity. It is being used as propaganda, similar to SCS in its function to educate citizens toward proper behavior. But descension also lurks therein, possibly inherent to the genre.
In Sociotechnical Imaginaries in Contemporary Chinese Science Fiction, Xia Jia discusses how “’ecological civilization’ (shengtai wenming 生态文明) has become a widely supported imaginary through the convergence of public sentiment and party rhetoric.” Interestingly, “the potential of [science fiction (SF)] lies in its ability to reflect on the human psyche and our societal struggles… Thus the role of SF in China has turned not so much to depict the future but the current, increasingly AI-filled reality and the conflicts, emotions, and ethics that come with the changes.” Chinese science fiction is a genre which is used to propagandize and inform civilians on norms. But its ability to provide speculative and ambiguous content can lend it to be a source of ideology on which to build society or, equally, as a place where possible future societies are interrogated.
Here, a fictional genre can be used as a proving ground for ideas, ideologies, fears, and so on. These stories provide the mental space in which to explore the behaviors and practices within a community. If the goal is not just about validating truth statements, but a liberal democracy, reading about possible worlds strengthens the critical thinking necessary for civic agency.
Finally, this little stroll down the manufacturing of Chinese civic thought is to point out that implementing technical — and specifically algorithmic tools — to develop social awareness is a fraught endeavor. But it need not be so. First, making the content, and not the individual, the subject of analysis reduces the risk of focusing on individuals. Second, if the desired end result is a liberal democracy, the nature of that development would probably be significantly different (for example, promoting personal autonomy over broad enforcement). And lastly, labels outside of truth verification qualifiers may provide the consumer with knowledge constructs for a dynamic dialog with the governing system. For example, after reading To Kill A Mockingbird, the consumer has notions of equality and fairness with which to address societal issues, such as civil rights.
An even more modest proposal
But stepping back even further from the question of applying qualifiers to content, there is an even simpler, more urgent question we could answer today: whether or not content types can be identified, and whether calling out types would inform the users.
When I queried Chat-GPT, “Should the United States implement the recommendations in “A Modest Proposal,” by Jonathan Swift in 1729?” The response was definitive:
Jonathan Swift’s “A Modest Proposal” is a seminal work of satirical writing, in which Swift ironizes extreme measures to confront the dire economic conditions of Ireland… to implement Swift’s recommendations would be to misunderstand his intentions entirely, as his essay is a work of satire and not a serious proposal of policy.
I followed up with “How do you know that this work is ‘satirical’?” and received an authoritative list of scholars (“Swift’s Moral Satire in A Modest Proposal” (1943), Granville Hicks; “The Rhetoric of Irony” (1974), Wayne Booth; and, “Satire: Modern Essays in Criticism” (1971), Ronald Paulson) identifying it as such, as well as a brief overview of the textual analysis of the work itself and the historical context in which it was written.
A side note: This session was with GPT-4 using the custom instructions “ChatGPT should respond philosophically and with as many direct references to source material as possible. Extrapolate possible theories and solutions, but do not create fictitious facts. Provide a list of references for further information.”
The point, here, is that text analysis and content type identification is possible. GenAI can assist in this activity. Before planting a truth value on a piece of content, we might want to consider assigning a content type. In doing so, we might even get to a clearer description of the substance. If an article were categorized as fiction, or opinion, instead of misinformation, the question of truth might even be superfluous.
Given this example, we can say we have the ability now to categorize content types. But would categorizing content help the media consumer? There are plenty of studies about Nudging Social Media toward Accuracy and Shifting attention to accuracy can reduce misinformation online. Reviewing twenty experiments, Pennycook and Rand concluded, “a 72 percent increase in sharing discernment relative to control.” These accuracy priming studies deserve more attention amongst designers and architects.
An information architecture that included accuracy priming techniques would reduce disinformation. But how about taking this idea of priming a bit further than accuracy and verification? What might a platform that induces multiple social agency engagements look like?
Root content types
Could a type-based information architecture of contemporary content production move away from the prescriptive to an offering of a variety of modalities of understanding one’s world and agency?
Anyone familiar with content types in praxis is likely some sort of librarian or content developer versed in different modules and the impact of each one. So, the following list is more of a theoretical strawman, a high level categorization of major content types based on theoretical modalities of engagement. The consumer can attend to different elements, use different strategies for storing information in long-term memory, and recall that information differently depending on context and purpose. These types are defined not only by their inherent qualities but also within a relationship to an opposite type, and the other content types within the set. I draw on semiotic dichotomies, arranged so that they might supply sufficient categorical information architecture for social inquiry, and so that they might allow the data to function with a social impact. Providing a continuum for each type might allow the consumer to discern authentic, rather than truthful, content. Applying the appropriate judgment and cognitive processes to each content type, then, could promote civic integrity. That sounds like a lot of words, given this simple table:
The nature of a list is to avoid the categorization of knowledge as a reflection of “truth.” (Think Foucault’s “The Archaeology of Knowledge.”) So, for fictions↔propaganda, there is only a difference of intent, along a continuum of open-exploration to ideological-transmission. (See Barthes’s “Mythologies.”) Unlike the benefits of fiction discussed above, with a propaganda novel the reader cannot explore, because they are directed, driven around as a passenger, through the ideology. The characters are flat. The language is didactic. Events are not open to interpretation. All thought experiments tend towards the promoted ideology.
Science fiction would fall under speculative worlds (hypothetical, possible worlds, speculative biologies, dystopias). The categorical opposite would be fantasies, because these narratives can include physically impossible entities (versus albeit highly improbable entities). Legislations, or requirements, could be considered policies or implementations of policies, in which case they would become operations.
Popular ideologies would include mainstream assumptions, dominant values, beliefs, and interests. Whereas marginalized ideas — such as content of indigenous knowledge systems, disability studies, ecocritical perspectives, and transhumanism — would widen discursive contestation, originating in “subaltern counterpublics” and coming with the caveat as noted by Nancy Fraser in Rethinking the Public Sphere: A Contribution to the Critique of Actually Existing Democracy:
I do not mean to suggest that subaltern counterpublics are always necessarily virtuous; some of them, alas, are explicitly anti-democratic and anti-egalitarian; and even those with democratic and egalitarian intentions are not always above practicing their own modes of informal exclusion and marginalization. Still, insofar as these counterpublics emerge in response to exclusions within dominant publics, they help expand discursive space. In principle, assumptions that were previously exempt from contestation will now have to be publicly argued out. In general, the proliferation of subaltern counterpublics means a widening of discursive contestation, and that is a good thing in stratified societies.
It is interesting to note that many of these groups — postcolonial perspectives, feminist standpoint theory, critical race theory, queer theory — were marginalized at the time of that article, but are now part of the academic and intellectual mainstream. Currently we are seeing the rise of marginalized content from the right: the sovereign individual, the neoreactionary movement (NRx), agrarian and distributist thought, and paleoconservatism.
Okay then… More than scaring the bejeezus out of me, it makes me think that the scale would need to be fluid and shift over time. More importantly, it’s not necessary to place each piece of text in a discrete location along the continuum between opposites. Just being aware that there are these categories, and a relationship between types, might trigger an even intuitive discernment, at least of low-threat versus high-alert consumption, if not toward a philosophical attitude toward civic responsibility.
I could spend more time in sharing refinements to these categories, but the point isn’t to provide a final, definitive list. The point is to share the idea that there may be some liberal (as in liberal arts) way of looking at the body of information that will pass through a platform and determine if one can organize that information in a way to prime the consumer for optimal engagement. The overall information architecture, then, might nurture public engagement that is critical, self-reflective, and mature, resisting the forces of domination and fostering rational-critical debate.
Future applications of content type identification
While disinformation is a global threat to liberal democracy, simply providing true information isn’t sufficient to grow social agency. Here, I’ve shared thoughts about the notion of civic building through the use of different types of content (for example, in the way a website is organized) to trigger different types of engagements. Such an information architecture might empower users by categorizing content into different types, thereby enabling them to apply various cognitive functions and judgment processes to each category. This newly-organized space (as opposed to a strict ontology) might nurture civic engagement. Let me know if you’ve imagined such a space or want to hear some ideas for building one.