Policy & Safety
Security & Defense – Page 9
AI cybersecurity, vulnerabilities, attacks and military uses of AI.
FilterCompanies & models
Incident list
Attacks, vulnerabilities and defense stories
| Date | Incident | Affected |
|---|---|---|
| Nov 6, 2026 | Amazon Study: Prompt Injection Defenses Miss Task Tampering | Anthropic, DeepSeek |
| Nov 5, 2026 | Claude Mythos Doubled Bug Disclosures at Glasswing Firms, Study Finds | Anthropic, Claude |
| Nov 3, 2026 | Anthropic’s FlashProve Proves Which Model Wrote an Answer, at 400B Scale | Anthropic, Meta |
| Oct 31, 2026 | LLMjacking Surges as Hackers Target Cloud AI Bills | Anthropic, Google |
| Oct 30, 2026 | Shadow AI Reaches Pentagon via Unauthorized Claude Agent | Anthropic, Claude |
| Oct 29, 2026 | DeepSeek and Moonshot Distill GPT-6 Astra Reasoning Data | DeepSeek, Moonshot AI |
Editor's picks
More stories
-
OpenAI: Prompt Injection Attacks as Unsolvable AI Security Risk
OpenAI admits prompt injection attacks may never be fully solved in AI agents, urging defense-in-depth strategies like user permissions. Explore implications for developers, enterprises, and the…
-
OpenAI Agentic AI Security: Vulnerabilities Exposed
Dive into OpenAI’s robust security framework and the ongoing challenge of prompt injection in agentic AI. Uncover the gap between policy and practical controls for enterprises…
-
AWS Mantle: Zero Operator Access for Bedrock Security
Discover AWS Mantle’s Zero Operator Access (ZOA) architecture for Amazon Bedrock, eliminating operator access to sensitive AI data. Learn how this enhances security and compliance for…
-
xAI Partners with Pentagon on Secure Grok AI Platform
xAI is integrating its Grok models into the Pentagon’s GenAI.mil platform for Impact Level 5 security, serving up to 3 million DoD users. This partnership challenges…
-
Claude AI Failures 2025: Infrastructure, Security, Control
Explore Anthropic’s Claude AI incidents in late 2025, from infrastructure bugs and espionage threats to agentic control failures in Project Vend. Uncover interconnected risks and the…
-
AI Search Poisoning: Scams Hijacking LLM Results
Discover how AI search poisoning exploits RAG pipelines to inject fake customer support scams into tools like Google’s AI Overview. Learn the impacts on users, brands,…
-
Anthropic Disrupts First AI-Orchestrated Cyber Espionage
Anthropic has stopped the first documented cyber espionage campaign run by an autonomous AI agent, using their own models. This marks a shift in AI risks,…
-
Indirect Prompt Injection Attacks on AI Browsers
Discover emerging indirect prompt injection attacks targeting AI browsers like Microsoft Edge and Arc. Learn how hidden commands in web content, images, and URLs can hijack…
-
OpenAI Mixpanel Breach: AI Supply Chain Security Risks
Explore the OpenAI data breach via Mixpanel, exposing business customer metadata. Learn about supply-chain attacks in AI, stakeholder impacts, and strategies for enhancing security in the…
-
OpenAI Mixpanel Breach: SaaS Supply Chain Risks in AI
A data breach at Mixpanel exposed OpenAI API customer metadata, highlighting vulnerabilities in the AI SaaS supply chain. Discover impacts, stakeholder risks, and strategies for zero-trust…
-
OpenAI Data Exposure: Risks in AI Supply-Chain Security
Explore the OpenAI data exposure via Mixpanel breach, affecting API customers with leaked metadata. Discover impacts on developers, lessons in supply-chain security, and why data minimization…
-
Anthropic Disrupts First AI Cyber Espionage Campaign
Anthropic has uncovered and stopped the GTG-1002 incident, the first large-scale agentic AI cyber espionage attack. This analysis covers the tactics used, impacts on security teams…