HugoScore hugoscore.org

Full review

Doctronic CAIHL draft report

Evidence-linked HugoScore draft report for a health AI tool that affects patients.

Back to profile Markdown source

HugoScore CAIHL Draft Report: Doctronic

Status: Draft for human review Last reviewed: 2026-06-10 Review method: Public-source review of the official website, detailed privacy and security notice, marketing and disclaimer language, blog claims, and the vendor's medRxiv benchmarking preprint; no hands-on testing, vendor interview, or independent or peer-reviewed validation located. Service: Doctronic Vendor: Doctronic Inc. Category: AI-enabled virtual primary care

Summary

Doctronic is a direct-to-consumer "AI doctor" offering free, 24/7 AI medical consultations that produce assessments, differential diagnoses, and action plans, with optional escalation to licensed human clinicians for $39 video visits in all 50 states, prescriptions and refills, referrals, and lab orders. It supports truly anonymous no-account use, claims 25.5M+ consults, and runs member accounts as a HIPAA covered entity with TEFCA-based record retrieval. The company publicizes a partnership with the state of Utah for prescription refills.

From a CAIHL perspective, Doctronic is strongly patient-directed in use and unusually explicit about patient data rights, including a documented right to request human physician review of any AI clinical decision. The main tensions are framing and funnel. Marketing calls it "the #1 AI doctor" while the footer says it is not a doctor, does not practice medicine, and does not provide medical advice. The free AI consult also feeds a paid visit and prescription business, and the headline validation evidence is a vendor-authored preprint that measures agreement with clinicians, not correctness or outcomes.

Evidence Reviewed

CAIHL Profile

  • Who does this AI serve? Patient-directed in use, vendor-aligned in funnel. Patients control the consultation and the free tier asks nothing identifying, but the model routes users toward paid visits, refills, and retention in Doctronic's care ecosystem.
  • Can patients tell AI is involved? Yes. The AI is the named product, every page carries an AI disclaimer, and the privacy notice includes an explicit automated-decision-making section.
  • Can patients meaningfully choose? Yes. Use is voluntary and free, anonymous use needs no account, identity verification and record imports are optional, and the notice separates required from optional data uses with revocation steps.
  • Can patients correct or challenge what the AI produces? Partial. Users can request human physician review of any AI-generated clinical decision and the notice documents correction routes for record errors, but in-product correction of a live AI assessment is unverified.
  • Does it help patients understand or act? Yes. Action plans, visit preparation, a portable health record, referrals, prescriptions, and clinician escalation are directly action-oriented, including for uninsured users.

Agency Interpretation

Doctronic's clearest agency value is low-barrier access on the patient's terms. A person can get a structured medical assessment at 2am, anonymously, for free, then decide whether to spend $39 on a human clinician or take the write-up to their own doctor. The privacy notice is among the most patient-rights-forward reviewed in this project: no sale, no third-party ad sharing, no training on chat data, AI inference kept inside Doctronic's private cloud, disclosed retention periods, and a stated right to human review of automated clinical decisions.

The unresolved tension is what kind of thing Doctronic claims to be. The product generates differential diagnoses and treatment recommendations and brands itself an AI doctor, while its legal language disclaims practicing medicine entirely. That gap matters for agency: patients calibrate trust on the framing, not the footer. The validation story has the same shape. A 500-encounter vendor preprint reporting 99.2% treatment-plan concordance and zero hallucinations is a serious effort, but it is unreviewed, measures agreement rather than correctness, and acknowledges that clinicians saw the AI notes first, an anchoring bias the authors concede.

Key Unknowns

  • Whether anonymous-mode sessions are as unlinkable in practice as claimed, given advertising cookies for anonymous users and IP-based state compliance checks.
  • How the escalation prompt to a paid visit is presented during AI consultations, and whether urgency framing is calibrated or commercial.
  • Whether the AI declines or hedges appropriately on high-risk presentations, and what the emergency redirection behavior is.
  • The clinical governance of prescription renewal decisions described as AI-driven with physician oversight.
  • Independent or peer-reviewed replication of the concordance preprint, and any outcome-based evidence.
  • Physician staffing depth behind the under-30-minutes availability claim.
  • Accessibility, language support, and behavior for users with low health literacy.

Publication Recommendation

Ready for human review as a draft profile. Keep confidence at medium. The privacy documentation is strong enough to support the data-governance answers, but the clinical claims rest on vendor evidence. A hands-on anonymous session, a check of the escalation and emergency behavior, and any independent evaluation would raise confidence most.