Research Project

Who is responsible when AI misdiagnoses a patient?

Active

governancebenchmark

Project Team

Dr. Suvrankar Datta

2025-Present · JIPMER, AIIMS Delhi, Ashoka University

CRASH Lab treats governance as a design problem, not a compliance afterthought. The goal is to define what responsible clinical AI deployment should require before systems scale.

Related projects

Published

Can AI read a chest X-ray as well as a radiologist?

A landmark benchmark testing frontier multimodal AI against board-certified radiologists on complex, multi-finding radiology cases.

radiologybenchmarkmultimodal-AIchest-xray

RSNA 2025 — Cutting Edge Oral Presentation

View research →

Seeking Collaborators

Why do AI diagnostics fail in Indian clinical contexts?

Systematic evaluation of how frontier AI models perform on clinical data from Indian healthcare settings versus the Western datasets they were trained on.

benchmarkIndia-healthcarecomputer-vision

View research →

Published

Can AI autonomously write radiology reports that radiologists trust?

Multidimensional evaluation framework for autonomous optimization pipelines in personalized head CT and chest radiograph report generation.

report-generationNLPradiology

RSNA 2025

View research →