Bias, consent and equity audit layer with provenance reports.
Bias, consent and equity audit layer with provenance reports.
GLIO-ETHIC is one of 11 subsystems that compose AURORA-GLIO. It is independently usable but shares the same patient representation, provenance log and governance layer as every other AURORA module.
Equity is not an audit at release. It is a loss inside the training loop. GLIO-ETHIC is the layer that enforces that contract — equity-as-loss across stratifications defined in the module RFC, with consent metadata as a first-class object and provenance reports generated alongside every model card.
Failures here block releases. There is no exception path for marketing reasons. The threshold is part of the RFC, citable and revisable in public.
GLIO-ETHIC can be installed standalone, or pulled in as part of the full AURORA-GLIO stack. Both ship the same code and weights, signed end-to-end. Pilot partners have access today; public alpha opens Q4 2026.
pip install aurora-glio[ethic] aurora glio.ethic audit \ --model aurora-glio/diag-v1 \ --strata ancestry,sex,age
GLIO-ETHIC moves faster when the right people are in the room. Whether you carry a service line, a registry, or a few quiet weekends — there is a way in.
Sponsor GLIO-ETHIC for your unit. Co-design the surface, the audit log and the override semantics with the team that built it. Pilot sponsors keep a council seat for the duration.
Run a research question on GLIO-ETHIC. Federation means you can train across populations you cannot see alone. Whitepapers carry co-authorship, not acknowledgements.
GLIO-ETHIC ships starter tasks tagged 'good-first-issue' on the pilot repo. SDKs in Python, TypeScript and Rust at public alpha. MIT, no CLA.
AURORA is in private pilot today and opens to the world at public alpha in Q4 2026 under MIT. If you carry the weight of these diseases — as a clinician, scientist, builder, patient or advocate — there is a seat at the table.