AI Adoption Science

A field needs a record before it needs a theory

A research series tracing one question: how do you add governance to a reasoning system without degrading the reasoning? Each claim is tagged by evidence — observed, corroborated, peer-reviewed, or industry-validated — so you can see exactly how far each one has earned its keep.

The Series

Each tag marks the evidence maturity of the article’s central claim in our taxonomy — not a peer-review status of the article itself. See the taxonomy →

I
O-1 Phenomenon

The Governance Paradox

Why heavy governance can make AI look safer while reasoning worse.

Corroborated ~1,600 words Read Article →
II
O-2 Hypothesis

The Sycophancy Risk

Why AI models tell you what you want to hear — and why that is dangerous.

Peer-reviewed Science 2026 Read Article →
III
O-4 Principle

Trust Elasticity

Why AI governance should tighten only when performance slips.

Corroborated ~1,700 words Read Article →
IV
O-5 Pattern

Bifrost: Cross-Session Governance

Tiered memory and hierarchical summarization that carry trust across sessions without carrying the noise.

Industry-validated ~1,900 words Read Article →
V
Meta-framework

The Classification Problem

A scientific taxonomy for distinguishing what is observed, corroborated, and actually proven.

In draft Coming 2026

The Taxonomy

76 statements. Sorted by how much we actually know.

76
Classified statements
41
Observed in the field
32
Corroborated by literature
3
Peer-reviewed or industry-validated
The Research Agenda

Only three statements have reached the top of the ladder. That near-empty summit is not a gap to hide — it is the agenda. Every claim still sitting at “observed” is an open invitation to a controlled study.

Industry-validated1
Peer-reviewed2
Corroborated32
Observed41

Explore the Constellation

Evidence Standards

“A claim that cannot be falsified is a slogan. We’d rather publish a small number we can defend than a large one we can’t.” How claims advance at Janus Labs

Observed → Corroborated

At least three independent observations, then external literature or a second source that aligns. Useful, not final.

Corroborated → Peer-Reviewed

A controlled study with falsification criteria, examined by independent reviewers. The hard gap in our current evidence base.

Peer-Reviewed → Industry-Validated

Independent teams reproduce the result in production at scale. Community acceptance, not a single lab’s claim.

Follow the Research