AI Medical Model for Clinical Reasoning Simulation

I investigated the problems that cause hallucinations in current medical LLMs and designed a medical model to reduce and detect them before generated output reaches a specialist physician. My research findings led to a patent application for a safe, comprehensive medical AI model intended to support physicians across all medical specialties.

Engineering Clinical Diagnosis

I defined a visual framework for the behavioural and engineering disorders that may limit the reliability of current medical AI models.

Clinical AI Disorders

Each term below describes a distinct reliability problem rather than a comparison between models.

Hallucinations

Generating medical information, facts or recommendations that are unsupported, inaccurate or not grounded in a reliable source.

Overconfidence

Presenting an answer with excessive certainty despite ambiguity, insufficient information or an inability to verify the conclusion.

Catastrophic Forgetting

Loss or displacement of previously learned medical knowledge when newer information or model updates alter existing representations.

Unsupported Outputs

Producing clinical statements without adequate evidence, source verification or a transparent basis for the answer.

Clinical Risk

The potential for unreliable AI output to mislead interpretation, influence decisions or compromise patient safety.

Current Model vs Proposed Direction

I developed this visual comparison to distinguish current model limitations from my proposed research direction.

Research Comparison

I structured this comparison to present current limitations and my proposed research direction without exposing protected implementation details.

Current Medical AI

  • Language prediction
  • Unorganised medical information
  • Weak source verification
  • Answers despite insufficient information
  • Potential clinical risk

Proposed Research Model

  • Clinical reasoning simulation
  • Organised medical information
  • Mechanisms to reduce catastrophic forgetting
  • Governance behavioural protocol
  • Continuous verification process
  • Patient-safety-oriented design
Validation Phase

CPIP–Python Code

I created a dedicated validation phase to test the behavioural protocol through Python-based evaluation while keeping the internal engineering implementation undisclosed.

CPIP_PYTHON_CODE / VALIDATION_PHASE