Market Pulse

AI Loss Events: The Three-Year Escalation in Frequency and Severity

This is a 22‑page board report analyzing 48 verified AI loss events from January 2023 to June 2026, mapped to the NIST AI Risk Management Framework and split into financial and non‑financial harm.

Core topics

• Rising AI‑incident frequency using AIID and OECD data.

• Shift from reputational issues to wrongful arrests, healthcare denial, deepfake fraud, and loss of life.

• Governance failures and tightening enforcement (EU AI Act, FTC, EEOC, courts).

• Contrast between cash losses and harder‑to‑price civil‑rights, privacy, safety and trust harms.

Structure

• Cover and scope summary with NIST taxonomy and event count.

• Executive Summary with key statistics and board messages.

• Methodology & Taxonomy, including causation categories and three‑level severity scale.

• Sections on Frequency, Severity/Causation Mix, and Financial vs. Non‑Financial Impact.

• AI Loss Event Register table of all 48 incidents.

• Observations, Conclusions & Next Steps, plus references and disclaimer.

Charts

• Annual AI‑incident bar chart (2019–2025).

• Exploitation‑window line chart (771 days to under 2).

• Causation and severity bar charts.

• Donut charts for quantified loss vs. non‑financial harm.

• GenAI fraud‑loss projection line chart to 2027.

• Governance‑gap bar chart for 2025 breaches.

Event register and board framing

• 48 entries with date, description, organizations, financial figure (if any), non‑financial impact, severity, and source.

• Seven causation buckets: security, fraud, reliability, bias, privacy, governance/regulation, hallucination/misinformation.

• Concludes AI loss is now a base‑rate enterprise risk, recommending a dedicated register, governance upgrades, and specific 0–12 month actions for risk, security, finance, legal and audit leaders.

You need to have an account in order to read more and download this report.