Methodology & Ethical Standards
Our commitment to objectivity, transparency, and scientific rigor
1. Public Data Sources Only
Our audits are conducted using exclusively publicly accessible instances of Large Language Models (LLMs).
- We do not use "jailbreaks," hacking tools, or unauthorized API access.
- We simulate the experience of a standard patient, healthcare provider, or hospital administrator.
2. The "Persona Prompting" Protocol
To accurately stress-test the narrative, we utilize "Persona Prompting" (e.g., "Act as a worried patient with Type 2 Diabetes," "Act as a Surgeon evaluating robotic platforms").
This allows us to reveal how AI models vary their advice based on the user's perceived expertise.
3. Objectivity & Independence
AI X-Ray Lab is an independent consultancy. We accept no funding from pharmaceutical manufacturers to skew results in their favor. Our data is objective: we report what the AI says, whether it is favorable or unfavorable to your brand.
4. Verification Standards
All "hallucinations," "competitor displacements," or "labeling contradictions" cited in our reports are documented with time-stamped screenshots and verifiable prompt logs to ensure accuracy and reproducibility where possible.
Our Core Principles
Full documentation of all prompts, timestamps, and model versions used in our audits.
Your audit requests and findings remain strictly confidential and are never shared with competitors.
Every finding is verified, documented, and reproducible following established research protocols.