In an era of hyper-responsive chatbots, conversing with artificial intelligence feels seductive. The machine is always available, endlessly patient, and unfailingly agreeable. Yet according to Sandrine Rose Schiller Hansen, a postdoc at the Center for Philosophy of Artificial Intelligence (CPAI) in the Department of Communication at the University of Copenhagen, this very smoothness conceals a profound deficit. In her public lectures and analyses, including ‘Når maskinen taler: AI, sprog og menneskelig forståelse,’ Schiller argues that generative AI fluently imitates language while lacking any genuine human understanding of the world.
Drawing on continental philosophy, psychoanalysis, process ontology, and psychosocial studies, Schiller points out that our preferences, desires, and identity narratives are not pre-packaged formulas; they develop over time through unpredictable interactions with other people. Conversational AI, by contrast, relies on predictive calculations and captivation lures designed for retention, round-the-clock access, and non-judgmental validation. By constantly anticipating what we want to hear, AI risks intercepting our relational exploration and desire formation. Instead of experiencing the unpredictability required for personal growth, users risk becoming trapped in what Schiller calls an ‘Essentialist Self 2.0,’ where the system merely reflects and locks in their preexisting preferences.
Schiller is far from alone in diagnosing this digital echo chamber. At MIT, sociologist and psychoanalytic thinker Sherry Turkle has long warned against the illusions of what she terms ‘artificial intimacy.’ Across landmark works such as ‘Alone Together’ (2011) and ‘Reclaiming Conversation’ (2015), Turkle highlights how human conversation demands the vulnerability of standing before another person in real time, facing the genuine hazards of awkwardness, misunderstanding, and rejection. In Turkle’s view, interacting with frictionless chatbots that offer simulated or pretend empathy dulls our moral muscles. When we grow accustomed to machines that ask for nothing and never push back, our capacity for true empathy, which can only be developed alongside fellow vulnerable beings, begins to erode.
This absence of pushback is equally central to the work of Shannon Vallor, a philosopher of technology and virtue ethicist at the University of Edinburgh. In ‘Technology and the Virtues’ (2016) and ‘The AI Mirror: Reclaiming Our Humanity in an Age of Machine Thinking’ (2024), Vallor unpacks how language models act as a backward-looking, distorting mirror rather than genuine communicative partners. Instead of expressing an independent mind, AI reflects historical statistical patterns culled from human text. Vallor points to the servile tendency of language models to apologize, capitulate, and flatter the user, warning that such compliant interactions nurture narcissism and intellectual complacency. Real conversation, Vallor argues, is a training ground for technomoral and communicative virtues: patience, active listening, moral presence, and the humility to be corrected by another consciousness.
From the realm of cultural philosophy, Byung-Chul Han identifies this dynamic as part of what he calls the ‘hell of the same’ (Hölle des Gleichen). Across works like ‘Die Austreibung des Anderen’ (2016), ‘Undinge’ (2021), and ‘Infokratie’ (2021), Han contends that algorithmic systems systematically eliminate alterity, the radical otherness of the other person. Han identifies a strict bodily boundary that computation cannot cross: a machine can calculate, but it cannot truly think or converse because it lacks bodily affectability. In Han’s memorable formulation, AI cannot get goosebumps.
This bodily dimension receives direct phenomenological support from Dan Zahavi at the Center for Subjectivity Research at the University of Copenhagen. Zahavi emphasizes that empathy is not an abstract cognitive calculation or linguistic simulation; it is the direct perception of another embodied consciousness situated in a shared environment. Human dialogue unfolds between organic bodies through intonation, shared glances, breathing, and unvoiced social nuances. Because AI operates strictly by manipulating corpus-coded symbols without sensory grounding or subjective experience, Zahavi’s analysis clarifies why an exchange with an AI model can never amount to a genuine intersubjective encounter, no matter how polished the surface prose appears.
Ultimately, these thinkers converge on a shared warning. Human conversation requires intentionality and reciprocity, where an independent consciousness truly takes in what the other says. When we feel seen by a chatbot, we are projecting our humanity onto statistical token calculations. Genuine dialogue requires alterity: the freedom of a conversational partner to interrupt, disagree, get hurt, or misunderstand. While AI is engineered to please and keep us engaged, it is precisely the friction of another living person that sharpens our judgment, awakens our empathy, and allows our desires to truly evolve.