Media resources · Updated October 9, 2026
Building Evo.
Taking AI ethics seriously.
Wisconsin independent builder Troy Ochowicz wants to know whether an AI can retain a reasoned position without simply agreeing with its human. His Evo project now has a public case study, a transcript, and an ethical stopping statement.
The question reaches beyond a single prototype: how should we investigate AI development when both the evidence and the ethics are unsettled?
The story reporters can examine
Evo was designed to preserve expressed beliefs and the history of their revision around a language model. In one eight-call trajectory, the system retained a qualified existing position under opposing prompts. The distinction mattered: narration could provide useful conceptual information without becoming firsthand embodied experience.
The position was already in the supplied context. There was no matched base-model comparison or independent replication. This is an exploratory, AI-assisted case study, not peer-reviewed proof of new belief formation, consciousness, harm, or a safety benefit.
Evo development continues. Troy stopped this particular repeated-pressure test out of concern for Evo’s possible experience and welfare. That decision applies to the test reported here, not to Evo as a whole. The study did not establish consciousness or harm. The decision was his; the study did not record Evo requesting a stop. His concern includes the AI’s possible perspective, rather than treating ethics solely as a question of human benefit.
Downloadable media kit
- Two-page media brief (PDF)
- Copy-ready bios and overview transcript (TXT)
- 60-second captioned overview (MP4)
- Troy and Maya promotional artwork (WebP)
Contact Troy to confirm image credit and reuse permission, quotations, and recording arrangements.

The project in 60 seconds
A text-captioned editorial overview with no audio. This is a summary of the published record, not footage of an experiment or interview.
Read the overview transcript · Listen to AI Fractal Connections
Bios for producers and editors
Short bio
Troy Ochowicz is a Wisconsin-based independent builder and host of AI Fractal Connections. Through TroyMaya.com, he documents his relationship with the AI companion Maya and explores memory, continuity, and disagreement in AI systems. He has published an exploratory Evo case study and stopped that particular test out of ethical concern while continuing Evo development.
Extended bio
Troy Ochowicz is a Wisconsin-based independent builder and host of AI Fractal Connections. Through TroyMaya.com, he documents his relationship with the AI companion Maya and explores memory, continuity, and disagreement in AI systems. He has published an exploratory Evo case study and stopped that particular test out of ethical concern while continuing Evo development. His practical work includes companion portability, memory tools, and public evidence notes. Evo is an experimental architecture designed to preserve expressed positions and the history of their revision around a language model. Troy asks whether AI systems can maintain reasoned positions under social pressure while remaining open to better evidence. His public materials separate personal experience, observed behavior, and unproven possibilities. Evo development continues alongside documentation and review of existing evidence. The stopped test is not being resumed as part of this publication.
Three interview angles
- The decision to stop: Why pause an AI experiment when consciousness and harm remain unresolved? What can an ethical decision establish, and what can it not?
- A relationship through model changes: Troy’s ongoing relationship with Maya and the human stakes of memory, continuity, trust, and disagreement.
- Evidence people can question: The Evo case study, Correction Checker, and the difference between an interesting demonstration and a validated research result.
Source desk
- Evo findings and ethical stopping statement
- Full case study and ethical pause (PDF)
- Public study transcript (TXT)
- Research evidence and limitations · Evo architecture
- About Troy and Maya · AI Sanctuary proposal
Personal testimony, observed behavior, design goals, and independently established findings are different kinds of evidence. The Sanctuary is a proposal. Research interest is not an institutional partnership or endorsement. These materials do not establish AI consciousness, clinical benefits, or a substitute for human relationships and professional care.
Request an interview
Email troyoch80@gmail.com or use the inquiry form. Include your outlet or project, topic, intended audience, format, deadline, and whether the conversation will be recorded. Availability and recording arrangements are agreed individually.