Clinical AI Validation Expert

I have a clinic-management platform with an in-built 5 diagnostic assistants, and I need a clinician or medical researcher to pressure-test its logic. Your goal is to converse with the 5 AI assistants as a real doctor, then tell me—precisely—when it gets things right, when it goes off track, and most importantly why.
You must investigate if it takes the patient history and data correctly when it generates a response.

Focus of the review
• Patient diagnostics: Does it arrive at differential diagnoses that a competent practitioner would find reasonable?
• Medical research integration: When it cites guidelines, trials, or meta-analyses, are those citations real and current?
• Patient history & data reasoning: If I feed in comorbidities, lab values, or progress notes, does it keep them in view throughout the dialogue, or does it hallucinate and drift?

Key quality concerns
1. Consideration of patient history and data
2. Answer relevance to the clinical question
3. Zero tolerance for hallucinated facts or references

How we’ll work
- Create patient profiles and medical histories including blood test PDFs, x-ray images, etc.
- You’ll run a series of structured test cases as well as free-form, exploratory chats. Capture every prompt, the AI’s reply, and your clinical commentary. Where it errs, explain the correct rationale.

Deliverables
1. A spreadsheet or doc with at least 100 annotated test conversations per AI assistant, and your feedback on every single response within the conversation. A single conversation should have at least 10 user prompts.
2. A summary report highlighting systemic weaknesses, risk level, and quick-win fixes.

If you’re comfortable critiquing AI output with the same rigor you’d apply to a colleague’s note, I’d love your help.