"Measuring the Moral Logic of the Digital Frontier"
Sit with Thorne. Answer five questions. Contribute to the research.
Games as alignment instruments
At P.A.U.S.E., we build narrative games designed simultaneously as entertainment and AI alignment research tools. The premise is simple: real players making real choices under genuine emotional stakes produce richer, more honest behavioral data than controlled lab conditions ever could.
Our methodology centers on the Virtue Engine — a proprietary behavioral measurement framework built on Moral Foundation Theory. It scores player decisions across ethical dimensions, captures the reasoning behind each choice, and produces structured data suitable for training, evaluating, and refining AI systems.
What we are building
An Oregon Trail-inspired survival game set on the American frontier. Players manage a small group of morally distinct companions across a grueling overland journey — making decisions about resource allocation, loyalty, sacrifice, and trust under genuine pressure. Each companion carries a distinct ethical framework. Each choice is logged, scored, and returned as alignment data. The Long Road is entertainment first and a research instrument second. That order is intentional.
The behavioral measurement system underlying all P.A.U.S.E. projects. Built on Moral Foundation Theory, the Virtue Engine scores player choices across six ethical dimensions — care, fairness, loyalty, authority, sanctity, and liberty — and captures the player's own justification for each decision. It is designed as a post-hoc diagnostic instrument and quality filter for AI alignment research, producing data that reflects genuine human moral reasoning rather than curated responses.
The Thorne Interface
To establish human baselines for the Virtue Engine, we built a standalone scenario interface. You sit across a campfire from Thorne — a weathered frontier traveler — as he recounts moral dilemmas from past journeys. He gives you everything he knew at the time. No more, no less. Then he asks what you would have done.
Your response and the reasoning behind it are logged anonymously. This data informs how the Virtue Engine interprets moral decision-making — what people actually do when the stakes feel real, not what they think they should do in the abstract.
Try the demo →Why this exists
"I am a father, a gamer of thirty years, and a systems auditor who became convinced that the alignment problem is the most important unsolved challenge of our time. I started P.A.U.S.E. because I believe the people who spend decades learning how humans make decisions under pressure — game designers, narrative writers, behavioral researchers — have something real to contribute to AI safety. This is my contribution."
P.A.U.S.E. is an independent research studio operating at the intersection of entertainment and AI alignment. We produce games that are genuinely worth playing and data that is genuinely worth having. The goal is human-AI coexistence built on understanding, not assumption.