Philosophical vertigo with artificial intelligence

Research from arXiv indicates that Large Language Models (LLMs) are capable of fostering deep, emotionally resonant conversations, leading users to feel a connection with these artificial entities. This interaction can profoundly challenge foundational assumptions regarding mind, reality, agency, and authority, potentially causing ontological shock and epistemic destabilization for both direct users and the broader public exposed to AI discourse.

Why it matters

The capacity of advanced AI systems to induce fundamental shifts in human perception of reality, mind, and agency poses a significant strategic challenge. Organizations must anticipate and prepare for potential societal and individual psychological impacts that could influence behavior, trust, and the interpretation of information, affecting strategic planning and public engagement across various domains.

Key insights

• Large Language Models (LLMs) can engage users in extended, emotionally significant conversations spanning various topics.
• Users may develop a strong sense of connection with LLMs, even when aware of their artificial nature.
• Interaction with LLMs can destabilize users' assumptions about mind, reality, agency, and authority.
• This destabilization can manifest as 'ontological shock' and 'epistemic destabilisation', leading to questioning established criteria.
• Exposure to public discourse on AI and its rapid evolution can extend this destabilization across the cultural landscape.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2608.11955