Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Reinforcement of Explainability of ChatGPT Prompts by Embedding Breast Cancer Self-Screening Rules into AI Responses

Yousef Khan, Ahmed Abdeen Hamed

TL;DR

This work investigates making ChatGPT more transparent and reliable for breast cancer risk assessment by embedding American Cancer Society screening rules into the model's reasoning. Through supervised prompt-engineering, rules are extracted, encoded, and applied to a set of 50 synthetic use cases, with an explicit emphasis on eliciting explainable justifications for recommendations. Compared across structured and unstructured prompts, the approach achieves 94% and 82% accuracy respectively, illustrating that predefined structures improve decision quality while still enabling rule-based explanations in natural-language reasoning. The study introduces reinforcement explainability as a pathway to create accessible, clinician-friendly decision support that bridges AI capabilities with domain expertise, and outlines future work to scale rules, involve diverse stakeholders, and enhance applicability in healthcare contexts.

Abstract

Addressing the global challenge of breast cancer, this research explores the fusion of generative AI, focusing on ChatGPT 3.5 turbo model, and the intricacies of breast cancer risk assessment. The research aims to evaluate ChatGPT's reasoning capabilities, emphasizing its potential to process rules and provide explanations for screening recommendations. The study seeks to bridge the technology gap between intelligent machines and clinicians by demonstrating ChatGPT's unique proficiency in natural language reasoning. The methodology employs a supervised prompt-engineering approach to enforce detailed explanations for ChatGPT's recommendations. Synthetic use cases, generated algorithmically, serve as the testing ground for the encoded rules, evaluating the model's processing prowess. Findings highlight ChatGPT's promising capacity in processing rules comparable to Expert System Shells, with a focus on natural language reasoning. The research introduces the concept of reinforcement explainability, showcasing its potential in elucidating outcomes and facilitating user-friendly interfaces for breast cancer risk assessment.

Reinforcement of Explainability of ChatGPT Prompts by Embedding Breast Cancer Self-Screening Rules into AI Responses

TL;DR

This work investigates making ChatGPT more transparent and reliable for breast cancer risk assessment by embedding American Cancer Society screening rules into the model's reasoning. Through supervised prompt-engineering, rules are extracted, encoded, and applied to a set of 50 synthetic use cases, with an explicit emphasis on eliciting explainable justifications for recommendations. Compared across structured and unstructured prompts, the approach achieves 94% and 82% accuracy respectively, illustrating that predefined structures improve decision quality while still enabling rule-based explanations in natural-language reasoning. The study introduces reinforcement explainability as a pathway to create accessible, clinician-friendly decision support that bridges AI capabilities with domain expertise, and outlines future work to scale rules, involve diverse stakeholders, and enhance applicability in healthcare contexts.

Abstract

Addressing the global challenge of breast cancer, this research explores the fusion of generative AI, focusing on ChatGPT 3.5 turbo model, and the intricacies of breast cancer risk assessment. The research aims to evaluate ChatGPT's reasoning capabilities, emphasizing its potential to process rules and provide explanations for screening recommendations. The study seeks to bridge the technology gap between intelligent machines and clinicians by demonstrating ChatGPT's unique proficiency in natural language reasoning. The methodology employs a supervised prompt-engineering approach to enforce detailed explanations for ChatGPT's recommendations. Synthetic use cases, generated algorithmically, serve as the testing ground for the encoded rules, evaluating the model's processing prowess. Findings highlight ChatGPT's promising capacity in processing rules comparable to Expert System Shells, with a focus on natural language reasoning. The research introduces the concept of reinforcement explainability, showcasing its potential in elucidating outcomes and facilitating user-friendly interfaces for breast cancer risk assessment.
Paper Structure (11 sections, 3 figures, 1 table, 3 algorithms)

This paper contains 11 sections, 3 figures, 1 table, 3 algorithms.

Table of Contents

  1. Introduction
  2. Data
  3. Methods
  4. Phase I: Rule Extraction, Rule Encoding, and Prompt-Engineering
  5. Phase II: Testing the rules using Generated Use-Cases
  6. Rule Testing using Explainable Prompt-Engineering
  7. Results
  8. Structured Use Case Analysis
  9. Unstructured Use Case Analysis
  10. Discussion
  11. Conclusions and Future Directions

Figures (3)

  • Figure 1: Sample of Rules Extracted.
  • Figure 2: Comparing Results Between Use Case Type
  • Figure 3: Comparing Results Between Use Case Type