AI and clinical safety
Early access · updated June 27, 2026
This Policy describes how Illumate’s AI features should and should not be used. This is a draft; the final version requires legal and clinical review.
Intended use
AI features are intended to assist with documentation: draft notes, non-diagnostic administrative summaries, preparation prompts, and search over your session history. All outputs are drafts.
Excluded use
AI features are not intended for emergency response, crisis triage, suicide or self-harm risk assessment, diagnosis, treatment planning, medication decisions, or any autonomous clinical decision. Illumate is not a medical device.
Human review
You must review, edit, and approve AI outputs before using them in a clinical record or sharing them with any third party. Each AI output is labelled as a draft; there is no automated send to an EHR without review.
Limitations
AI outputs may contain errors or “hallucinations”, be incomplete, or be inapplicable to a particular case. Guardrails reduce but do not eliminate out-of-scope output.
Auditability
Actions on content and AI outputs are logged; where applicable, human approval of an output is recorded for oversight purposes.
Incident reporting
Report unsafe or incorrect AI outputs to hello@illumate.me so we can improve guardrails and Service quality.
Illumate is in early access, so these documents may be updated as the service evolves. The provider and contact details are set out in the sections above.