AI Mediation in the War Over “Truth”

The Medium is a Battlefield (We Are Strong)

The emergence of AI mediation systems like the "Habermas Machine" reveals a deeper crisis in our digital age. While these systems promise to help humans find common ground, they enter a world where the very infrastructure of truth and knowledge is contested territory. Digital platforms aren't simply neutral spaces for deliberation and finding common ground, but they are a battleground where different groups fight to control how reality itself is understood and verified.

This core tension extends beyond what we might call the mediation paradox - where our ideological divisions require computational assistance beyond human comprehension. Recent research shows AI systems can successfully mediate group deliberation where direct human interaction fails. Yet this apparent success masks a more fundamental question: How can AI mediate between groups who inhabit different belief systems and epistemic architectures, each with their own ways of determining and verifying truth?

The materiality problem in our digital age presents a fundamental paradox. While material mediations shape human subjectivity – our interactions with objects, tools, and physical infrastructure – these mediations have now become sites of active ideological warfare. Digital platforms aren't neutral spaces where different viewpoints naturally converge; they are contested battlegrounds where reality itself is actively constructed and fought over.

Consider Musk's transformation of Twitter into X: this isn't merely a change in ownership but a deliberate reconfiguration of an epistemic architecture.

By altering verification systems, content moderation, and algorithmic priorities, this move demonstrates how control of digital infrastructure means control over how truth is verified, distributed, and legitimized. Different ideological groups don't just hold different beliefs - they inhabit entirely different digital ecosystems with distinct ways of knowing and experiencing reality.

This transforms our question about AI mediation. It's no longer simply about whether AI can help bridge differences but whether AI mediation systems can function effectively in a world where the very infrastructure of knowledge is contested territory. Can AI help reconcile fundamentally different epistemic architectures, or will it reinforce these divisions? We face a race between technological development and philosophical understanding and a challenge to the possibility of shared truth in digitally mediated space.

Moving forward requires reconceptualizing both human agency and democratic deliberation within this contested digital landscape. Rather than seeing AI mediation as either a threat or a solution, we must understand it as part of knowledge production and verification infrastructure. This suggests developing hybrid systems that aren't just about combining human judgment with computational assistance but about creating new epistemic architectures that can acknowledge and potentially bridge these fundamental divisions in how different groups understand and verify reality.

This contested digital landscape demands new philosophical frameworks that go beyond traditional understandings of materiality and mediation. We need new theories to explain how different systems of knowledge and truth-verification emerge, compete, and might coexist in digital spaces. These frameworks must recognize a crucial fact: the infrastructure of knowledge itself - from social media platforms to AI systems - has become the primary battlefield in ideological struggles.

As Mill observed about the ideological conflicts of his time, between opposing parties, there exists a 'bellum internecinum' (a war of mutual extermination), where one side sees its opponents as 'beasts' while the other condemns its rivals as 'lunatics' (Mill, 1840, p. 405). But today, this war isn't just about conflicting beliefs - it's about controlling the very architecture through which reality is constructed and verified.

The challenge isn't simply technical or philosophical but fundamentally about power and control over the architectures of truth. While AI offers tools for mediating disagreement, we must always ask: Who controls these systems? Whose epistemological framework do they privilege? How can they function when the very ground of shared reality is contested? We need ways to harness AI's mediating capabilities while ensuring they don't become another weapon in epistemic warfare.

Moving forward, we must face a brutal truth at some point: no amount of code will make people join hands and sing kumbaya. Technology won't save us from our very human selves. The sooner we abandon the fairy tale, the sooner we can deal with the reality of human conflict in all its sharp-edged glory is an ever-present aspect of our “reality”.

Instead, we must develop systems that can operate effectively within and across competing epistemic architectures while maintaining democratic legitimacy in the struggle over meaning and power in the context of politics. This means designing AI systems that mediate between different viewpoints and acknowledge the fundamental problem that we have different ways of knowing and verifying truth, and sometimes people simply refuse to play fair. The struggle to control meaning will always be present, even in moments where consensus is achieved.

The stakes here transcend traditional questions of democratic deliberation and human agency.

Some individuals express their disdain for others' feelings by displaying flags with offensive messages like "fuck your feelings." 

How should we respond to such provocations? Is consensus even possible in this context? Admitting this reality problematizes the Habermasian project, which relies on the notion of rational discourse in the public sphere. To address these challenges, it's important to look beyond Habermas for a more comprehensive understanding of the role of recognition, social struggles, and the psychological dimensions of human development in shaping public discourse. This is where Axel Honneth's critique of Habermas becomes particularly relevant.

We are witnessing a struggle over the infrastructure through which reality is constructed and understood. How we approach AI mediation in this context will determine how we resolve political disagreements and who controls the architecture of knowledge in the digital age. Our challenge is to ensure these emerging systems serve democratic flourishing rather than becoming tools for epistemic domination.

This challenge resonates with Axel Honneth's critique of Jürgen Habermas' theory, as presented in "The Theory of Communicative Action" (Habermas, 1981). Honneth believed that Habermas' work fell short in addressing the importance of recognition, social struggles, and the psychological dimensions of human development in shaping society and social progress.

Honneth proposed to correct these shortcomings by reconnecting critical social theory with anthropological materialism, as evidenced in his seminal works such as "The Critique of Power: Reflective Stages in a Critical Social Theory" (Honneth, 1991), "The Struggle for Recognition: The Moral Grammar of Social Conflicts" (Honneth, 1995), and "Disrespect: The Normative Foundations of Critical Theory" (Honneth, 2007).

As we navigate the complexities of AI mediation in the context of competing epistemic architectures, Honneth's insights remind us to consider the role of recognition and social struggles in the pursuit of democratic legitimacy and the construction of shared knowledge.

Honneth’s ideas, as outlined in the aforementioned works, provide a valuable framework for understanding and addressing the challenges posed by the digital age in light of Habermas's proposed solutions and AI's increasing influence in shaping our social and political realities.

Bibliography:

Habermas, J. (1981). The Theory of Communicative Action, Volume One: Reason and the Rationalization of Society. Beacon Press.

Honneth, A. (1991). The Critique of Power: Reflective Stages in a Critical Social Theory. MIT Press.

Honneth, A. (1995). The Struggle for Recognition: The Moral Grammar of Social Conflicts. Polity Press.

Honneth, A. (2007). Disrespect: The Normative Foundations of Critical Theory. Polity Press.

Mill, J. S. (1840). Coleridge. In J. S. Mill, Dissertations and Discussions: Political, Philosophical, and Historical. (Vol. 1, p. 405). John W. Parker.

Tessler, M. H., Bakker, M. A., Jarrett, D., Sheahan, H., Chadwick, M. J., Koster, R., Evans, G., Campbell-Gillingham, L., Collins, T., Parkes, D. C., Botvinick, M., & Summerfield, C. (2024). AI can help humans find common ground in democratic deliberation. Science, 386(6623), eadq2852. https://doi.org/10.1126/science.adq2852