Medical AI Safety Research Team: Build the "Crash-Test Lab" for Clinical Chatbots

About the Project

We are launching a 12-week research initiative at the BRITE Institute to address a critical, overlooked safety issue how medical AI systems react when faced with missing, buried, or contradictory patient information. We are building an "Adversarial Agent"—a robotic stress-tester for medical AI—to measure when these systems are safe and when they are prone to dangerous assumptions.

Qualifications

• Commitment to a 12-week project timeline.
• Mission-driven, collaborative, and interested in AI safety and healthcare quality.
• Skills in a mix of technical (Python/API), academic (Data analysis/Writing), and clinical (Medical expertise) backgrounds.

Why Join Us

• Impact: Help set the safety standards for the future of AI in medicine.
• Authorship: All key contributors will be credited as co-authors on our final academic research paper.
• Growth: Gain hands-on experience building an LLM-based testing infrastructure and navigating the clinical-AI safety landscape.