Research explained

Healthcare breach-cost figures need context before they become an AI-risk claim

A reported average breach cost describes a study population, not the expected loss from every AI incident.

Published 29 Oct 2026 · Material update 31 Jul 2025 — last time the underlying evidence changed · Evidence: medium

What happened

Managed Healthcare Executive’s July 2025 coverage of IBM’s breach-cost study reports an average healthcare breach cost of $7.42 million. The study described in the article covers organizations affected between March 2024 and February 2025.

That is a historical study statistic. It is not a price tag for a particular hospital’s AI deployment, and it does not mean every breach in the sample was caused by AI.

For decision-making, separate the observed population, cost definition, and incident period from broader commentary about AI adoption. Local exposure, data sensitivity, and response capacity still need to be assessed independently.

Read the original source.

What we know

Evidence type
Reporting on a historical cost study

What remains unknown

Open question
The reported average does not predict the loss from an individual incident.

Source references