Editorial desk
AI Incidents Editorial
AI Incidents Editorial is the publishing identity for AI Incidents. It is a desk, not a person: no named author, no biography, no professional certifications.
Articles published under this byline are researched from primary sources — vendor and project documentation, published standards and specifications, research papers, and measurements published by whoever took them — drafted with AI assistance, and edited against those cited sources before publication. Nothing here is based on first-hand testing in a private lab, and any figure that appears is attributed to the source it came from.
Corrections go to hello@aiincidents.org. More detail is on the about page and the editorial disclosure.
Posts (16)
- analysis
Self-Driving Car Accident Causes: What Crash Data Shows
Sensor failures, software edge cases and fault attribution behind self-driving car accident causes, plus what NHTSA Standing General Order data shows.
- news
AI Incident Database Comparison: Which One to Use
AIID, OECD AIM, AIAAIC, AVID and MITRE ATLAS all answer to the name AI incident database. What each one records, and which to reach for when.
- news
AI Incident Reporting Requirements: Who to Notify
One AI incident can start four legal clocks at once. The EU AI Act, GDPR, SEC and sector regulators each want a different notice on a different deadline.
- analysis
Real World AI Failure Examples That Triggered Lawsuits
Real world AI failure examples, from a wrongful arrest in Detroit to a chatbot advising glue on pizza, and the court rulings and lessons they produced.
- tooling
Best AI Monitoring Tools 2026: LLM Observability Compared
A practitioner's guide to the best AI monitoring tools for 2026, covering LLM observability platforms, ML drift detection, pricing, and how to choose.
- explainer
Prompt Injection Attack Explained: How Attackers Hijack LLMs
A technical explainer of how prompt injection attacks work, the direct and indirect attack classes, real consequences, and what defenders can do.
- tooling
Deepfake Detection Tools Review: Platforms Compared
A deepfake detection tools review of Reality Defender, Intel FakeCatcher, Pindrop Pulse, Sensity AI, and Amber Authenticate, with a buyer decision matrix.
- news
Reconstructing an Incident Timeline From Primary Sources
A vendor advisory, a CVE record, a regulator filing and a researcher's blog post all date one event differently. How to reconcile them into a timeline.
- news
An Incident-Response Playbook for AI Systems
Generic IR runbooks assume the failing component is a server you can patch. AI incidents add a model whose behavior you can't fully explain.
- news
Anatomy of a Vendor Advisory: Reading What Isn't Said
Vendor advisories from AI providers follow a recognizable shape. Knowing what to look for, and what is left out, turns marketing into usable signal.
- news
AI Taxonomy: Incident, Vulnerability, Disclosure, Misuse
A working taxonomy that distinguishes AI incidents, vulnerabilities, disclosures, and misuse by impact, weakness, publication, and intent.
- news
NVD CVE Entries for ML Libraries: What Fields Mean
NVD entries for torch, transformers, vLLM, and LangChain are not written for ML engineers. How to read the fields and what to do when metadata is thin.
- news
Reading a Model Card for Security Signals
Model cards are written for researchers, not defenders. What to read first, in what order, to judge a model's security posture and disclosure practice.
- news
Source Verification Tiers: Vetting an AI Incident Claim
A five-tier source ladder for verifying AI security incident claims, with evidence thresholds and examples of claims accepted or rejected.
- news
Why We Don't Do Attribution Speculation on AI Incidents
Attribution is the slowest, hardest, most consequential call in incident reporting. Here's the policy that keeps us from getting it wrong.
- news
How We Log AI Security Incidents: Our Methodology
The methodology behind AI Incidents — how we verify sources, date-stamp claims, and decide what's news vs noise in the AI security incident beat.