Description: A purported AI tool using LLMs was deployed to classify Veterans Affairs contracts as expendable based on limited text and simplified criteria. The system allegedly produced hallucinated values and flagged critical healthcare and research services for cancellation. Reportedly, at least two dozen flagged contracts were later terminated.
Entités
Voir toutes les entitésPrésumé : Un système d'IA développé et mis en œuvre par Department of Government Efficiency , DOGE , Sahil Lavingia , Custom LLM-based contract classifier ("Munchable" tool) et OpenAI large language models, a endommagé Department of Veterans Affairs (VA) , Veterans , Veterans receiving care through the VA , VA clinical and research staff et VA contractors.
Systèmes d'IA présumés impliqués: Custom LLM-based contract classifier ("Munchable" tool) et OpenAI large language models
Statistiques d'incidents
ID
1103
Nombre de rapports
2
Date de l'incident
2025-03-18
Editeurs
Daniel Atherton
Rapports d'incidents
Chronologie du rapport
As the Trump administration prepared to cancel contracts at the Department of Veteran Affairs this year, officials turned to a software engineer with no health care or government experience to guide them.
The engineer, working for the Depar…
The Department of Government Efficiency (DOGE) developed an artificial intelligence (AI) tool to identify unnecessary contracts at the Department of Veterans Affairs (VA), intending to cut costs. However, the AI system showed significant te…
Variantes
Une "Variante" est un incident de l'IA similaire à un cas connu—il a les mêmes causes, les mêmes dommages et le même système intelligent. Plutôt que de l'énumérer séparément, nous l'incluons sous le premier incident signalé. Contrairement aux autres incidents, les variantes n'ont pas besoin d'avoir été signalées en dehors de la base de données des incidents. En savoir plus sur le document de recherche.
Vous avez vu quelque chose de similaire ?
Incidents similaires
Did our AI mess up? Flag the unrelated incidents
Incidents similaires
Did our AI mess up? Flag the unrelated incidents