Autonomous AI
Also called: autonomous diagnostic AI
A system that produces a clinical result without a clinician interpreting the underlying data, currently authorized in a small number of narrow indications.
Autonomous AI makes the determination itself. The best-established examples are retinal screening devices that return a diabetic retinopathy result without any clinician viewing the images.
Authorization requires substantially stronger evidence than assistive tools, typically including prospective validation with the intended non-specialist users in the intended setting.
Where This Gets Misread
Autonomy shifts liability in a way the field has not fully worked through. When no clinician interpreted the study, responsibility for a missed diagnosis sits with the deploying organization and the manufacturer rather than with a reader. That is precisely why the evidence bar is higher, and why the category remains small.