COVID-19 Detection and Prognostication Models Allegedly Flagged for Methodological Flaws and Underlying Biases

January 1, 2020

A review of COVID-19 detection and prognostication algorithms from 2020, including deployed models, found that none were suitable for clinical use due to methodological flaws and underlying biases. Lacking external validation and unspecified data sources/model training details were common issues. These findings underscore the importance of safe and secure AI practices in developing reliable predictive models.

Join us at Project Cerebellum, where we are working on the HISPI TAIM to Govern, Map, Measure, and Manage incidents like this one, contributing to harm prevention efforts for trustworthy AI. To get involved, visit JOIN US.

Matched TAIM controls

Suggested mapping from embedding similarity (not a formal assessment). Browse all TAIM controls

MEASURE 2.6 — similarity 0.712, rank 1. TAIM detail and related incidents →
MEASURE 3.1 — similarity 0.676, rank 2. TAIM detail and related incidents →
MEASURE 1.1 — similarity 0.670, rank 3. TAIM detail and related incidents →

Alleged deployer: mount-sinai-hospital, unknown
Alleged developer: icahn-school-of-medicine-researchers, unknown
Alleged harmed parties: covid-19-patients, covid-19-healthcare-providers

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/535

Data source

Incident data is from the AI Incident Database (AIID).

When citing the database as a whole, please use:

McGregor, S. (2021) Preventing Repeated Real World AI Failures by Cataloging Incidents: The AI Incident Database. In Proceedings of the Thirty-Third Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-21). Virtual Conference.

Pre-print on arXiv · Database snapshots & citation guide

We use weekly snapshots of the AIID for stable reference. For the official suggested citation of a specific incident, use the “Cite this incident” link on each incident page.