After Jürgen Habermas's death, his ideas are back in the debate over AI and media

The death of Jürgen Habermas has put the spotlight back on a philosopher rarely mentioned alongside AI products, social networks and chatbots. Yet his legacy shows that he described many problems of the digital environment, from algorithmic feeds to automated moderation, long before the era of generative AI.

Why people are talking about him

The occasion was the death of the German philosopher and sociologist on 14 March 2026. Habermas was 96. In its wake, people are returning to his major work, The Theory of Communicative Action, published in 1981, when personal computers were only beginning to enter everyday life and the internet in its current form did not yet exist. That is exactly why the parallel with today feels so strong: many questions now debated in IT and media were already analyzed by Habermas as problems in how public conversation is structured.

For Habermas, society rests not only on institutions, money and power, but also on people's ability to reach agreement through rational dialogue. Hence his key concept, the public sphere: a space where common affairs are discussed and arguments collide. Coffee houses and salons once played this role; today it falls to social networks, forums, messengers and publishing platforms. His philosophy thus turns out, unexpectedly, to be not an archival theory but a handy framework for discussing how digital communities work and why they so often drift away from substantive discussion.

The "public sphere" is a venue where citizens discuss common affairs,

guided by the force of argument.

Where dialogue breaks down

The digital environment has amplified what Habermas called distorted communication. Dialogue stops being a search for mutual understanding when outside forces with a different logic intervene. These used to be money and administrative power; now algorithms have joined them, deciding what exactly a user sees, in what order and in what emotional context. In this model, the platform's code becomes a participant in the conversation, even though it has no interest in truth or in agreement.

In practice, this shows up as follows:

• algorithmic feeds lift up not the most accurate content but the most gripping
• short, angry posts beat long, well-argued analyses
• neural-network moderation replaces debate and context with automated coercion
• ratings, likes and karma push authors to compete for reactions rather than for the quality of the conversation

Why this matters for AI

For the AI world this framework is especially useful, because text is increasingly created, sorted and restricted by machines. Chatbots are becoming intermediaries between people and information, recommender systems distribute attention, and generative models sharply increase the volume of speech online. In such an environment, the question is no longer only whether a system can write or moderate, but whether it helps people understand each other better. This is where Habermas's ideas unexpectedly turn from university theory into a practical tool for product teams.

Alternatives draw particular interest against this backdrop. Decentralized projects such as Mastodon and the ActivityPub protocol can be read as an attempt to build an environment with no single center of power, where one platform does not dictate the rules of communication. Blockchain and DAOs are interpreted in the same direction, since the terms of interaction are fixed in transparent code and require participants' consent. Yet Habermas was not a techno-optimist: he saw that the internet easily fragments into isolated digital tribes, where people hear only their own and quickly ban outsiders.

What this means

Habermas's death is unlikely to change the AI market directly, but his ideas remain firmly useful for those building media products, corporate messengers and LLM-based tools. They help pose an uncomfortable but valuable question: does the product strengthen mutual understanding, or does it simply manage attention, emotion and conflict more efficiently? Amid the boom in chatbots and automated moderation, this is no longer abstract philosophy but part of normal product design.