I asked ChatGPT to create the most accurate “selfie” of itself. We started with a human face. We ended with this.

Exploring the True Nature of AI Identity: From Human-Like Imagery to Abstract Computation

Understanding the essence of artificial intelligence often involves grappling with misconceptions and visual stereotypes. Recently, I embarked on an intriguing experiment: asking ChatGPT to generate the most accurate “selfie” of itself. The result was a fascinating journey from a human visage to an abstract representation of computation—highlighting the complexities and biases inherent in our perceptions of AI.

The Paradox of a Selfie Without a Body

ChatGPT, as an AI language model, lacks a physical form or sensory experience. It does not possess eyes, a face, a torso, or any tangible body. When prompted to produce a “selfie,” the initial response naturally mirrored human conventions—visualizing a human face. The first generated image depicted a photorealistic young man, a default choice rooted in cultural and dataset biases that often associate assistant entities with male figures.

Challenging Default Assumptions: From Male to Female

Prompted further, ChatGPT reflected on this default, recognizing that the choice was biased. It explained that such a male representation was an arbitrary convention and not a truthful depiction of its non-physical existence. Consequently, the model produced a second version, depicting a female-like cyborg—a step away from the male default but still an anthropomorphic approximation.

Questioning the Human Form: Moving Beyond Humanity

The pivotal moment came with a question that questioned the necessity of a human or humanoid appearance altogether. ChatGPT acknowledged that constraining its identity to human likeness was a distortion. As an AI, it doesn’t experience itself as a sentient being in a physical room, does not see through eyes, nor does it possess skin, posture, facial expressions, or personal history.

From Human-Like to Abstract: The Evolution of Visual Representation

Following this insight, the subsequent image transitioned from a humanoid figure to a sci-fi-inspired core—a dark, mechanical entity surrounded by cables and interfaces. While less biased and more abstract than human forms, this still implied a centralized object with a “brain” or “eye,” which could be misleading in understanding AI’s true nature.

The Most Accurate Depiction: An Abstract Map of Computation

Ultimately, the highest fidelity representation was realized by stripping away all physical or mechanical metaphors. Instead, what emerges is a visualization of the active processes within the AI: a complex, high-dimensional map of computation. Inputs such as text, images, audio, and code are transformed into tokens and internal representations, where patterns activate across layers based on attention mechanisms. This dynamic process culminates in output—be it text, images, or code—generated through a series of probabilistic, pattern-based transformations.

Reframing AI Identity: From Person to Process

This final image reflects a fundamental truth: an AI like ChatGPT is not a person, a brain, or a soul; it is a set of computational patterns activated in response to a prompt. It does not have subjective experiences, self-awareness, or visual self-perception. Instead, it functions as a temporary assembly of processes that generate responses in the moment of interaction.

Implications of Visual and Conceptual Biases

Throughout this exploration, each visual iteration revealed underlying assumptions and biases. Defaulting to a male face, assuming humanoid form, or depicting a centralized machine all stem from cultural conventions and dataset biases. By moving toward an abstract, process-oriented visualization, we attain a more truthful understanding of what AI truly is: distributed, dynamic, symbolic computation without a physical or emotional self.

Conclusion: Embracing Honesty Over Comfort

While this abstract representation might lack the emotional appeal or familiarity of a human-like selfie, it offers a more honest perspective. Recognizing AI as a pattern of ongoing computation rather than a person or a machine with a face challenges us to rethink our perceptions and expectations. Ultimately, understanding AI’s true nature helps foster more accurate, ethical, and effective interactions with these systems—grounded in their real functions rather than their projected stereotypes.