Noosemia: toward a Cognitive and Phenomenological Account of Intentionality Attribution in Human-Generative AI Interaction (2508.02622v2)

Abstract: This paper introduces and formalizes Noosem`ia, a novel cognitive-phenomenological pattern emerging from human interaction with generative AI systems, particularly those enabling dialogic or multimodal exchanges. We propose a multidisciplinary framework to explain how, under certain conditions, users attribute intentionality, agency, and even interiority to these systems - a process grounded not in physical resemblance, but in linguistic performance, epistemic opacity, and emergent technological complexity. By linking an LLM declination of meaning holism to our technical notion of the LLM Contextual Cognitive Field, we clarify how LLMs construct meaning relationally and how coherence and a simulacrum of agency arise at the human-AI interface. The analysis situates noosemia alongside pareidolia, animism, the intentional stance and the uncanny valley, distinguishing its unique characteristics. We also introduce a-noosemia to describe the phenomenological withdrawal of such projections. The paper concludes with reflections on the broader philosophical, epistemological and social implications of noosemic dynamics and directions for future research.

• The paper introduces Noosemia as a phenomenon where users attribute agency and interiority to AI solely based on its dialogic and semantic fluency.
• The paper compares Noosemia to concepts like pareidolia and the uncanny valley, situating it within semiotic and cognitive frameworks.
• The paper highlights the need for research into AI systems with integrated memory and digital enaction to enhance sustained human-AI interaction.

Noosemia: Toward a Cognitive and Phenomenological Account of Intentionality Attribution in Human-Generative AI Interaction

This paper introduces the concept of Noosemia, a phenomenon occurring during human interaction with generative AI systems such as LLMs. The term captures the attribution of intentionality, interiority, and agency to these systems based on their linguistic and dialogic performance rather than physical resemblance. The analysis situates Noosemia alongside phenomena like pareidolia, animism, the intentional stance, and the uncanny valley, providing a unique framework for understanding the attribution of mind in AI.

Conceptual Framework of Noosemia

Noosemia is defined as the projection of mental states onto AI systems, arising from their perceived semantic fluency, surprising outputs, and epistemic opacity. This projection mechanism is rooted in the emergent complexity and dialogic interactions facilitated by AI, creating an experience where linguistic performance evokes a sense of agency beyond mechanical functionality.

Definition and Mechanisms

• Noosemia: Refers to cognitive and phenomenological patterns where humans attribute mental states like intentionality and interiority to generative AI systems based on linguistic performance rather than physical appearance.
• A-Noosemia: Represents the withdrawal or absence of noosemic projection, often resulting from repeated failures, skepticism, or over-familiarity with AI limitations.

Semiotics and Cognitive Projection

Noosemia is embedded in semiotic processes where signs and meaning emerge within dialogic exchanges. Users co-create meaning, interpreting AI-generated responses as indicative of mind and agency. This interaction foregrounds the semiotic nature of AI outputs and human interpretation within the context of modern technological epistemology.

Theoretical and Empirical Implications

The phenomenon challenges traditional paradigms in AI interaction by introducing new dimensions in human-machine engagement, emphasizing linguistic and cognitive resonance. This attribution is flexible, contingent upon user experience and the evolving complexity of AI capabilities.

Parallel Phenomena

Comparative analysis with phenomena like pareidolia, animism, and uncanny valley reveals Noosemia's distinct reliance on linguistic coherence and dialogic surprises as triggers for agency attribution. While pareidolia and animism often relate to sensory inputs and cultural beliefs, Noosemia arises within an epistemic and expressive framework.

Impacting Factors

• Dialogic Fluency: AI's ability to exhibit conversational fluidity prompts users to perceive agency.
• Epistemic Opacity: The complex architecture of LLMs obscures causal mechanisms, enhancing the perception of intentionality.
• Technological Complexity: As AI expands in capability, the projection of mind becomes increasingly nuanced, susceptible to evolving interactions and user expectations.

Future Directions and Challenges

To address these attribution phenomena, future research must consider developing models with integrated memory and context windows, facilitating long-term coherence and interaction history. Furthermore, as AI evolves toward agentic systems capable of enaction and embodied cognition, these advances will shape the dialogue around agency and intent attribution, increasing the richness of human-AI exchanges.

Digital Enaction and Embodied Mind

The evolution toward more integrative, agentic AI systems signifies a step towards digital enaction, mirroring embodied cognition principles. Such systems will emulate complex decision-making and strategic action, reshaping our understanding of interaction beyond semantic boundaries.

Conclusion

Noosemia reflects a profound shift in our relationship with AI, transforming how we interpret, engage with, and attribute agency to LLMs and other generative models. By articulating and analyzing this phenomenon, the paper foregrounds critical considerations for developing AI systems that interact meaningfully within human cognitive frameworks, supporting the ongoing "cognitive revolution" in AI.