Safety & Ethics
Shifa AI is designed to support, never replace, clinical judgment.
Human-in-the-loop
No automated clinical decisions; provider oversight required.
No PHI in pilots
Pilot data is de-identified or simulated.
HIPAA path planned
BAA & enhanced controls for enterprise deployments.
Our Safety Promise
Augment teams, keep clinicians in control, and default to caution.
View details
- Supports clinical judgment — alerts, not decisions.
- Clinician-in-the-loop by design — human review required.
- Built for coordination, not diagnosis.
- Conservative approach — when unsure, escalate.
Pre-HIPAA, Pilot-Ready
No PHI in pilots; controls in place; HIPAA path defined.
View details
- No PHI during pilots — de-identified/simulated data only.
- Technical controls — encryption, access control, audit logging.
- HIPAA path planned — BAA and enhanced controls at enterprise.
Clinician-in-the-Loop Safeguards
Human oversight with conservative triage and clear SLAs.
View details
- Conservative triage — escalate to humans when uncertain.
- Manual override — clinicians can reassess and resolve.
- Time-bound alert SLAs — tracked and auditable.
Transparency & Limits
Clear about what the platform does—and doesn't—do.
View details
- No diagnostic claims — not a medical device.
- Explainable alerts — show why a flag was raised.
- Human interpretation required — clinicians decide.
Data Privacy & Governance
Least-privilege design with redaction and retention controls.
View details
- De-identified data only in pilots.
- Logging with redaction; least-privilege access.
- Secure key handling and defined retention.
Roadmap to Compliance
Steps toward full HIPAA alignment and enterprise readiness.
View details
- BAA with cloud providers.
- KMS-backed encryption at rest.
- RBAC + audit log exports; periodic risk reviews.
Contact & Reporting
Questions or concerns? We're here to help.
safety@shifa-ai.com
- We reply to safety reports within 24 hours.
- Technical issues are addressed within 2 business days.