ADD Blogposts open a window to the work in and across the ADD-project. Meet our researchers from six Danish universities: Aalborg University, Aarhus University, Copenhagen Business School, Roskilde University, University of Copenhagen, and University of Southern Denmark. Read about their projects, activities, ideas and thoughts and maybe gain a new perspective on the controversies and dilemmas we face in digital democracy and how we can work to advance democracy in a digital age.
By Sine N. Just, professor of strategic communication at Department of Communication and Arts, Roskilde University, principal investigator of the ADD-project
Digital technologies have become ubiquitous mediators of public debate, shaping political exchanges of opinion as well as social processes of meaning formation in their image. In Controversial Encounters in the Age of Algorithms I argue that opposite forces of polarization and personalization converge to shape digital debate in ways that are detrimental to individual and collective decision-making. For reasons that I shall return to, I call the current constellation of digital public debate ‘controversial encounters of the second kind’, but what it really amounts to is a spectacle rather than a meeting, a display of controversy rather than an engagement with it.
Online controversies are circulations of digital content without any other goal than their continued circulation – despite individual motivations for engaging in such circulations, their collective effects amount to nothing but the further perpetuation of ever more of the same at ever-increasing speed. We saw a recent example of this in the rapid spread of variations of the ‘childless cat ladies’-meme in the run-up to the US presidential elections 2024. Appropriated from JD Vance’s 2021 statement that the US is being run by ‘a bunch of childless cat ladies’, in the fall of 2024 variations of the meme became widely circulated by supporters of the Harris-Walz ticket. For instance, Taylor Swift signed off her official endorsement of the Democrats’ candidates with the epithet, and producers of t-shirts and other merchandise were quick to spot the meme’s profitability.
What the heavy circulation of the ‘childless-cat-ladies’-meme illustrates is that, engaging and entertaining as such circulation may be, it does not have much effect. Or, as Jodi Dean (2019) puts it: “Circulation has eclipsed meaning. That something is shared online does not depend on what it means. It depends on its affective capacity: does the shared item manifest outrage: is it funny and diverting?” (pp. 331-332). This, Dean (2019) concludes, means reasons, ideas, and positions, are “…easily drowned out by outrage and puppies” (p.338).
I label these two forces ‘nasty conflict’ and ‘sweet consensus’, arguing that their combined effect is to inhibit our ability to actually encounter controversy. Sadly, when Democrats embraced the ‘childless-cat-ladies’-meme, they both confirmed the inutility of engaging with the opponent’s position (‘they’re just weird’, as Walz said) and the superiority of their own (‘look, the left can meme!’), thereby emptying out the space for mutual engagement and foreclosing the possibility of persuasion.
This kind of public debate emerged with digital technologies that are largely associated with web 2.0. Today, these technologies are being supplemented – if not, indeed, replaced – by generative AI. This raises the question of how the next ‘generation’ of digitalization (re-)shapes public debate. Are we witnessing an emergence of controversial encounters of a third kind? And if so, what does the new constellation look like? Before answering that question, let me explain what I mean by controversial encounters and establish ‘the first kind’ against which subsequent iterations are measured.
My concept of controversy is an amalgam of classical rhetorical understandings of ‘arguing from both sides of the case’ and modern science and technology studies’ take on ‘situations where actors disagree’ – and where the disagreement can neither be ignored nor easily resolved (Venturini, 2010). This is what facilitates the encounter; the understanding that disagreement is a precondition, not of isolation or, indeed, enmity, but of the concerted efforts to find a workable solution, acknowledging the contingencies and imperfections of any decision and remaining open to constant revision of it.
The notion of controversial encounters of the first kind is a conceptual ideal rather a historical reality. It is inspired by Jacques Ranciére’s (2004) dictum that a society’s capacity for disagreement is a measure of its plurality as well as its inclusivity. As such, controversial encounters of the first kind are characterized by ‘maximum capacity’ for open-ended meetings with others in the face of disagreement.
Controversial encounters of the second kind, to the contrary, display ‘minimum capacity’ for such meetings. In the current constellation, people have become unaccustomed to expressing their own opinions with persuasive intent and they have become wary of the persuasive attempts of others, leading to what I term the closing of the rhetorical mind – meaning, a situation where people no longer form opinions through the open exchange of ideas, but by means of algorithmically organized flows of data that are silently persuasive.
Here, I understand rhetoric as ‘reasoning with motive’, i.e., the merger of rationality and feeling to enable interlocutors to tune into each other. When our rhetorical minds are closed, we not only lose the ability to persuade others but also to be persuaded by them, shutting down the possibility that public debate can serve democracy by mediating between individual interests and public opinions.
Can generative AI help open the rhetorical mind and restore the democratic functions of public debate? This may seem like a rhetorical question that has an obvious answer and is only asked to feign audience involvement. However, it is, in fact, an open and contentious question that may have profound effects on how we understand persuasion, intention, language. The question is about the defining features of rhetoric, and I hope you will continue discussing it (with me) beyond the confines of this blog post.
In a tweet from March 2023, David Gunkel presciently indicates what is at stake in an early diagnosis of ‘the generative AI revolution’. ChatGPT, Gunkel observes, ‘writes without speaking’. Hence, generative AI makes clear that there is no original meaning behind any text, no ‘reality’ to which words strung together into meaningful texts refer. More puppies and outrage, then? A bottomless circulation of engaging but meaningless signs, ‘brain rot’ all the way down? This is one potential outcome: controversial encounters of the third kind as an extension and intensification of the spectacular outrage of the present into the future.
Still, generative AI, with its fundamental revelation of the contingency of meaning, might also lead us towards a deeper discussion of ‘the meaning of meaning’ and, less existentially, to a different form of automated persuasion. That is, further automation does not necessarily send people further down the vicious spiral of personalized polarization. It does not necessarily exacerbate social suspicion of overt persuasive attempts and leave individuals ever more vulnerable to the covert digital organization of persuasion.
We are in the middle of a transformation, and the options are still open; more of the same in ever more problematic ways may seem the more likely outcome – projecting the present course into the future. But transformations would not be transformative if they did not, in fact, offer opportunities to imagine and enact things differently. Beginning with alternative imaginaries, hope that generative AI might shape digital debate in directions that are more open to controversy as well as to encounter comes from two very different quarters.
On the one hand, artistic engagements with AI technologies reinvigorate art’s political potential for ‘defamiliarizing the familiar’, raising new questions of what we have otherwise come to take for granted (e.g., concerning human creativity and artistic originality), just as more explicitly political debates concerning artists’ (and other content creators’) intellectual property rights may help us ask old questions of what no longer can be taken for granted (i.e., reasserting the right to intellectual property and finding new ways of enforcing it).
On the other hand, technology developers are working to make generative AI more explainable and contestable, which are two prerequisites of engaging in controversy not only through but with digital technologies. Most notably, OpenAI suggests that its o1 model ‘thinks before it answers’, i.e., that it is able to reason, and Anthropic’s constitutive model for its AI assistant Claude indicates how stochastic probability can be guided by ethical principles.
While generative AI does not have motive (i.e., it is not independently intentional), its increasing aptitude for reasoning may give new rhetorical impetus to public debate, inviting controversial encounters in which humans and technologies work actively and openly together towards persuasive ends. Controversial encounters of this kind may (re-)open the rhetorical mind and, hence, enable frank discussions of how to live together with our social and technological differences intact. Today, few things are more important than that.
References
Dean, J. (2019): Communicative capitalism and revolutionary form. Millennium, 47(3): 326-340.
Rancière, J. (2004): Disagreement: Politics and Philosophy. Minneapolis: University of Minnesota Press.
Venturini, T. (2010): Diving in magma: How to explore controversies with actor-network theory. Public Understanding of Science, 19(3): 258– 273.