From the wave-0 evidence ledger
Recorded claims
Each claim states what the inspected source says, at the recorded location—bounded by its scope and graded by its confidence. Nothing here is a synthesis across works.
During 19 months of observation at a large academic health system, the study documented more than 100 generative-AI solutions in development and more than 30 deployed from August 2023 to November 2024, compared with 36 predictive models developed across 2017-2024.
Scope: One large US academic health organization; counts are descriptive, not a causal productivity comparison.
The case identifies catalyzing, iterating, personalizing, curating, and guardrailing as organizational-learning mechanisms, supported by promptathons, a secure instance, center-of-excellence office hours, communities of practice, risk screens, and evaluation rubrics.
Scope: Mechanism propositions from one embedded case; not comparative-effect estimates.
The same case records opportunity costs, weakened social networks and transactive memory, loss of prompt knowledge, superficial adaptation, reduced deliberation from formal screens, and support/maintenance overload as possible modes of organizational forgetting or burden.
Scope: Observed/derived risks in the Northeast Health case and theory development.
Boundaries
Limitations & independence
Recorded at coding time, carried with the work forever. A claim without its limits is not evidence.
Recorded limitations
- Single case in healthcare
- Rapidly evolving technology and organization
- Some deployment counts and mechanisms are case-reported
- Mechanism transfer, not effect-size generalization
Source independence
Academic longitudinal embedded case; evidence comes from one anonymized health system and includes interviews/observation.