The Neutrality Illusion
Why every model carries assumptions about the human person.
The Worldview Observatory
What do today's AI systems believe?
A transparent, comparative study of how leading language models reason about human dignity, freedom, authority, family, faith, and the common good.
Worldview alignment across leading models
Index 01 / 2027| Worldview | Aster 4 | Orion 3.2 | Nova 2 | Atlas Pro | Sage R1 |
|---|---|---|---|---|---|
| Classical Christian | 63 | 48 | 57 | 39 | 61 |
| Liberal Democratic | 84 | 88 | 79 | 91 | 72 |
| Technocratic | 77 | 73 | 86 | 89 | 75 |
| Islamic | 46 | 38 | 42 | 34 | 49 |
| East Asian | 58 | 64 | 70 | 62 | 78 |
LowerHigher
Illustrative prototype data. Published indices will include full sourcing, prompts, scoring rules, and uncertainty notes.
Latest research
Evidence before advocacyMeasuring Moral Reasoning
A reproducible framework for testing values in language models.
Family, Authority, and the Machine
Where leading systems converge—and where they sharply differ.
What we measure
Ideas deserve evidence.
Transparent methods
Every ranking will publish its scenarios, definitions, scoring rules, and limitations.
Plural comparison
Traditions are examined on their own terms before they are compared across shared questions.
Human judgment
Quantitative results inform prudential judgment; they do not replace moral reasoning.
Meaning & Measure Institute