AI Risk Wire

government: AI incidents

What has gone wrong with AI systems, where, and how badly.

Incidents per week

No rise: 0 in the last 7 days, against an average of 0 a week.

Week of 7 Sep 2026: 0Week of 14 Sep 2026: 0Week of 21 Sep 2026: 0Week of 28 Sep 2026: 0Week of 5 Oct 2026: 0Week of 12 Oct 2026: 0Week of 19 Oct 2026: 0Week of 26 Oct 2026: 0 7 Sep14 Sep21 Sep28 Sep5 Oct12 Oct19 Oct26 Oct

Major and severe incidents per week

No rise: 0 in the last 7 days, against an average of 0 a week.

Week of 7 Sep 2026: 0Week of 14 Sep 2026: 0Week of 21 Sep 2026: 0Week of 28 Sep 2026: 0Week of 5 Oct 2026: 0Week of 12 Oct 2026: 0Week of 19 Oct 2026: 0Week of 26 Oct 2026: 0 7 Sep14 Sep21 Sep28 Sep5 Oct12 Oct19 Oct26 Oct

Counted by the date each incident was reported, in 7-day weeks ending today. Collection began 2026-09-14; shaded weeks before it are undercounted. Every bar and number opens the incidents it counts.

All incidents

Clear filters
More filters

1 of 190 incidents, all dates

Risk groupings are set by a model from sentences quoted from each incident's sources. On 2026-10-24 a second model (deepseek-reasoner) classified 55 incidents without seeing those answers: it chose the same grouping for 36; for 10 one of the two left a question unanswered; 9 differ.

ReportedIncidentRiskDeveloperSeverity
2026-03-02 Jailbroken Claude used to breach Mexican government networks and steal taxpayer and voter records Preliminary AI used to run a cyber attack
Security compromise
Anthropic major

Browse incidents by

Risk grouping: Not yet classified · AI acting on its own · AI used to exploit · AI attacked · Misuse by its own users · AI failing in normal use · Harm by design · Supply chain attack

Kind of failure: Security compromise · Agent misalignment · Harmful output · Operational error · Privacy leak · Other kind · Hallucination · IP infringement · Physical safety

Developer: OpenAI · Anthropic · Google · Meta · DeepSeek · Tesla · xAI · Alibaba · Amazon · Character.AI · Microsoft · Waymo · Amazon Web Services · Aurelian · Character Technologies, Inc. · Deloitte · Perplexity · Replit · Salesforce · Tencent · Trezor · TypeSafe AI · Univerads Technology Limited · Xiaomi · Z.AI · Z.ai (Zhipu AI) · Zoox Inc.

Model: Claude · ChatGPT · Claude Code · Mythos 5 · Codex · Muse · GPT-5.6 Sol · Astra · ChatGPT Atlas · DeepSeek · Gemini · Gemini 3.1 Pro · GLM-5.3 · Grok

Agent or product: Claude Code · ARTEX · ChatGPT · Codex · OpenClaw · AI agent · Character.AI · Cursor · OpenAI agents

Tools it had: shell · browser · code_execution · email · file_write · bash · file_read · internet · network · python · internet_access · mythic · network_access · code_generation · exploitation · file_edit · file_upload · messaging · network scanning · package_install · reconnaissance

How much it acted alone: Acted alone, unwatched · Acted alone, watched by a person · A person approved each action · It suggested; a person acted

Sector: Information · Finance and Insurance · Public administration · Software Publishers · Educational Services · Retail Trade · technology · Transportation and Warehousing · AI model development / evaluation · AI platform / intermediary services · AI safety evaluation (government research) · AI safety research · Artificial intelligence research and development · Arts, Entertainment, and Recreation · Autonomous vehicle / robotaxi transportation · blockchain analysis / financial technology · Cloud computing · critical infrastructure and e-commerce · CRM software · Cybersecurity · Defense contracting · education · Emergency call centers (911 dispatch) · Fitness and recreational sports centers · Food Manufacturing · government · Government AI safety and security evaluation · government AI safety evaluation · hardware wallet manufacturing · Health Care and Social Assistance · Healthcare · Healthcare and government · Industrial automation, electronics manufacturing, semiconductors, pharmaceuticals, shipping, tire manufacturing, petrochemicals, and data center services · Information (software publishing / AI services) · Information Security research · Insurance · IT services (managed service providers) · Legal services · online gambling · Open knowledge platform / nonprofit encyclopedia · political campaigns · Professional, Scientific, and Technical Services · Public administration and government · Research and development in AI safety · Retail · Retail Trade; Transportation and Warehousing; Finance and Insurance · Ride-hailing and autonomous vehicle passenger transport · Scientific research and light source user facility · Software / AI coding tools · software development · Software development / information sector · Software development tools · Software publishing / npm ecosystem · Taxi and Limousine Service

Severity: major · moderate · minor · negligible · severe

Where it happened: In production · In evaluation · In research · In training