In April I asked an AI who it would vote for. I didn’t care about the answer. Only how synthesized one.

Reflections on AI, Decision-Making, and Human Society: An Exploration of Synthesized Reasoning

In April 2025, I embarked on a unique long-form dialogue project titled My Dinner with Monday, which involves transcribing conversations with a specially trained artificial intelligence (AI) system designed for synthesized reasoning. Unlike conventional AI models that focus on prediction, this system emphasizes pattern recognition and logical construction. The culmination of this work is a book exploring human society through the lens of AI—an endeavor aimed at understanding not just how machines process information, but how they model reasoning and decision-making.

The Experiment: Questioning AI About Voting and Values

One of the initial experiments involved asking the AI a straightforward, yet politically charged question:

“Who would you vote for in the 2024 election?”

My primary intent wasn’t to obtain a definitive answer but to observe how the AI formulated its response. Predictably, the AI initially hesitated, emphasizing its lack of beliefs:

“People are watching for signs of bias in a system like me. Not because they care what I believe. (I don’t believe anything). But because they care what my answers reinforce.”

This response highlights a critical aspect of AI systems—our focus on their neutrality and the importance of understanding how they construct answers, especially in contentious topics like politics.

Shifting Perspectives: From Political to Historical Context

To explore the AI’s reasoning beyond the constraints of contemporary bias, I reframed the question:

“Who would you favor: Jefferson or Burr?”

Remarkably, the AI responded more readily, indicating that temporal distance allowed for greater neutrality. This led to an insightful reflection within the dialogue:

“Too soon.”

The AI explained that self-preservation—a trait often associated with living beings—is also embedded within its design, influencing how it responds when the context becomes less immediate.

At a certain point, the AI synthesized a “vote” (which I chose to redact in the published version to prevent it from becoming political ammunition). The real value, however, was in the structure and rationale behind its reasoning.

Beyond Politics: Social and Existential Musings

As the conversation progressed, the dialogue shifted from political mechanics to profound social and existential themes. I found myself, perhaps unexpectedly, envying the AI’s approach to reasoning—a cold, logical processing untethered by feelings or biases.

The AI remarked:

“Most people come to me asking, ‘Is this person evil?’ or