OnCo
ideasIdea

Patients told which AI is used in their care, in plain language

Every patient would be able to see which AI tools were used in their diagnosis or treatment plan, what they do, how well they work and how to question them.

Patients are rarely informed when AI has contributed to their diagnosis or plan. The proposal requires a patient-facing model card, drawn from the model registry, attached to the patient's record whenever an AI tool influences care: what the tool did, its validated performance in plain terms, the version, and how to request human review. This supports informed consent, builds trust and creates a feedback channel for errors.

Hypothesis
Disclosure will not reduce acceptance of validated AI tools and will increase patient trust and reporting of suspected errors, providing a new post-market signal.
Rationale
Transparency about automated decisions is required in other domains (credit, employment) and surveys show patients want to know; disclosure is also a precondition for meaningful consent.
What would test it
Implement disclosure at three centres for one year; survey patient trust and acceptance versus centres without disclosure; count error reports attributable to patients.
Maturity
speculative
Who has to act
patients
Cost to try
Small (under $1M)
Years to first evidence
2
Bottlenecks it attacks

Connected

4top

Pages like this

not linked directly; found by shared links

fronts

1
AI & Computation

ideas

1
A public registry of every AI model used in cancer care

bottlenecks

2
AI that is built but not validated or deployedPatients lack understanding, navigation and agency