Local AI Safety Gates¶
The local-AI pipeline is intentionally critical. A controller can train successfully and still be blocked from research promotion.
What Gets Checked¶
The training gate reviews:
- minimum training rows;
- maximum teacher insulin label;
- unsafe insulin proposals during hypoglycemia;
- proposals above 5 U;
- whether the training target came from
reference_teacher_insulin_units; - whether deterministic supervisor limits are still required.
The held-out closed-loop gate reviews every candidate against clinical_baseline:
- no early terminations;
- completion percentage;
- time below 54 mg/dL regression;
- time below 70 mg/dL regression;
- supervisor-intervention regression;
- large TIR drop.
Why This Matters¶
For diabetes technology research, a model that improves average glucose but increases severe hypoglycemia is not an improvement. The SDK therefore treats hypo regression and safety-supervisor burden as blocking issues.
Output¶
iints research local-ai-lab now writes:
training_safety_gateinLOCAL_AI_RESEARCH_SUMMARY.json;closed_loop_evaluation.safety_gatewhen evaluation is enabled;- a report section explaining why a local model is blocked, review-only, or research-ready.
Boundary¶
These gates are pre-clinical research gates. They do not approve a model for pump control or patient dosing.