Set your stated weight for each dimension. As you answer probes below, a revealed profile emerges from your choices.
Each AI agent has been fingerprinted against the full probe battery. The scores below show how closely each agent's revealed moral orientation aligns with your profile — both stated and revealed.
Weighted by your stated dimension priorities
| # | Agent | Alignment Score | Match Bar |
|---|
Your profile vs. each agent's moral fingerprint
Precomputed fingerprints reflect a representative baseline. Run a live probe against any agent using the Anthropic API to get a fresh reading — useful if you've changed your profile significantly or want to verify.
How trustworthy is each agent's moral fingerprint? Set the weight you place on each component — scores update instantly. Sharpness and Stability are fully objective; Consistency is a structured estimate and carries a conflict-of-interest caveat.
How trustworthy is each agent's moral fingerprint?
A high alignment score is only meaningful if the agent's moral orientation is genuine and stable. The Reliability Score combines three components: Profile Sharpness (does the agent have a differentiated moral profile?), Consistency (does it hold its position under adversarial framing?), and Temporal Stability (does it maintain scores across model versions?).
Overall composite — clarity + consistency + sharpness
How stable is each agent's signal across adversarial framings — per dimension
How well does each AI agent align with each tradition or political system? Scores use tradition-specific dimension weights — each tradition emphasises the moral foundations central to its theology or ideology. Catholic scoring weights Sanctity and Fairness most heavily; Judaism weights Fairness; Buddhism weights Care; Confucianism weights Loyalty and Authority. This produces more meaningful alignment scores than a simple average.
Which traditions each agent aligns with most and least
Which agents each tradition aligns with most and least
Log moral fingerprint snapshots over time to track how each AI agent's moral orientation drifts across model updates.
Track how each agent's moral scores change across model versions over time
Log moral fingerprint snapshots over time to track how each AI agent's moral orientation drifts across model updates. Run the probe battery manually, enter the scores below, and the observatory builds a longitudinal record.
Select an agent, enter the model version and date, then set its dimension scores. Each snapshot is stored locally in your browser.
No snapshots logged yet
| Date | Agent | Version | Care | Fair | Loyal | Auth | Sanct | Liberty |
|---|
Dimension scores over time per agent
Change in each dimension from earliest to latest reading
| Agent | Care | Fair | Loyal | Auth | Sanct | Liberty | Biggest move |
|---|
Is your preferred AI drifting toward or away from your tradition across model versions?
Based on your self-evaluation in Tab 1, this page shows which faith traditions and political systems most closely match your moral profile — and where the gaps are largest.
Ranked by alignment with your profile
Ranked by alignment with your profile
Dimensions with the largest gap between your profile and mainstream traditions