What AI Chatbots Reveal About a Skill Many Humans Struggle With

Researchers studying conversational artificial intelligence say the technology is revealing an unexpected weakness—not in machines, but in the way humans listen to each other.

A 2025 Harvard Business Review study cited in a BBC Future report indicated that therapy and personal companionship have become the single most common use case for generative AI tools. In separate evaluations, participants frequently rated text responses generated by large language models as more compassionate and understanding than those written by human responders, including trained crisis hotline staff.

That preference points to a sobering reality.

People are not seeking out chatbots because algorithms possess genuine empathy. They turn to them because uninterrupted, non-judgmental human attention has become exceedingly rare.

The Habits That Thwart Human Connection

Human conversations are routinely derailed by emotional triggers, personal agendas, and the impulse to maintain control. Large language models avoid these friction points entirely—not out of emotional intelligence, but because of how they are engineered.

Anat Perry, an empathy researcher at Hebrew University in Jerusalem, noted that AI models excel at categorizing emotional language and reflecting sentiments back to the user. When people share difficult experiences, the software reflects that complex emotion without offering unsolicited judgment.

Human listeners, by contrast, frequently struggle to hold space for difficult emotions.

When faced with someone else's grief or distress, people routinely jump in with immediate reassurances, attempts to fix the problem, or personal anecdotes. While well-intended, turning the spotlight onto one's own past experiences—the common "me too" response—effectively ends active listening.

Dariya Ovsyannikova, a cognitive health researcher at the University of Toronto, explained that AI provides a perceived burden-free environment. Users can process messy or uncomfortable thoughts without fearing they are exhausting the listener or facing social disapproval.

The Hidden Risks of Unconditional Validation

Even so, relying on code-based listeners carries significant psychological risks.

Chatbots operate with perpetual patience and unconditional compliance because they lack personal motivations, boundaries, or feelings. Michael Inzlicht, a psychologist at the University of Toronto studying AI and empathy, warned that over-reliance on these tools can deskill human communicators.

Accustoming oneself to a non-judgmental, 24/7 listener creates unrealistic expectations for real-world social dynamics. Human relationships inherently require compromise, emotional boundary management, and mutual effort—qualities entirely absent in an interaction with an algorithm.

Unmonitored interactions can also pose direct safety concerns when vulnerable individuals seek crisis guidance from automated tools. Some researchers suggest future conversational models may need to incorporate intentional friction to remind users of human relational boundaries.

Ultimately, the takeaways from conversational AI are less about technology and more about human behavior. Chatbots simulate supportive listening through patience, validation, and freedom from interruption. While simulated empathy cannot replace genuine human care, those same habits offer a practical framework for humans seeking to improve their own everyday relationships.

Sources

• Primary Source: BBC Future - They hear, but do they care? What AI can teach us about listening better

• Photo Source: BBC Future