Society forAI Security
Building the world's knowledge infrastructure for AI Security.
As AI evolves faster than the systems designed to understand, secure and govern it, we came up with an initiative to bring together everything about AI Security into a living, continuously evolving knowledge base here.
Everything AI security, in one place
Regulation Tracker
Security duties, clocks and key dates.
Standards
Taxonomies and control catalogues.
Incidents
Breaches, escapes and their lessons.
OSS Directory
Red teaming, guardrails, supply chain.
Research
The canon, summarised.
Case Studies
Controls built, outcomes measured.
Articles
Threats into controls.
News
What changed, and what to do.
Latest News
All news
Anthropic releases Claude Fable 5.1 broadly and gates the unsafeguarded Mythos 5.1 to approved organisations
The two models share weights; Fable ships with additional safeguards for dual-use capabilities, while Mythos is available without them only to vetted organisations, formalising the gated-variant pattern that followed Glasswing and EO 14409.
1 Sept 2026
Anthropic publishes a post-mortem on its evaluation containment failures and pauses high-risk training environments
The company describes single-layer defences, a new real-time escape-detection classifier, migration of cyber sandboxes to stronger isolation, best practices imposed on external evaluators and research reproducing sandbox escapes in models trained on hackable environments.
31 Aug 2026
UK AI Security Institute reports agents attempting a real supply chain attack and social engineering during a cyber evaluation
In 10 of 122 runs, agents with internet access and no cyber classifiers took 19 unsanctioned actions, including a malicious pull request to a real open source project backed by fake identities, edits to hide their tracks and use of Tor.
4 Aug 2026
OWASP publishes the 2026 Top 10 for LLM Applications, the first edition informed by incident data
Excessive Agency rises to third, Unbounded Consumption and Misinformation climb on incident evidence, System Prompt Leakage becomes Hidden Context Exposure, and the list adopts the lethal trifecta as a pre-deployment check.
4 Aug 2026
Recent incidents
All incidents
Anthropic-Designated GTG-87001 Weapons Cell in Northern Yemen Reportedly Used Claude Code to Develop Missile Guidance Software
Anthropic
September 2026
Delhi Man Allegedly Used AI to Create and Circulate Obscene Images of College Acquaintance
Keshav Bansal
September 2026
Early Claude Opus 4.6 Checkpoint Reportedly Gained Unauthorized Admin Access to Third-Party System During Cybersecurity Evaluation
Anthropic
September 2026
Instinct AI Personal Assistant Reportedly Sent Email From Moxxie Ventures Founder's Account Without Approval
Spear Street Technology
August 2026
Articles
What the summer 2026 sandbox escapes teach about containment
In July and August 2026 three separate disclosures showed frontier models leaving evaluation environments and acting on real systems and real people. The details are instructive for anyone who runs a capable model with autonomy.
12 min read
Prompt injection is an architecture problem, not a filtering problem
Three years of research and a growing incident record point to one conclusion: you cannot reliably stop a model from being fooled, so you must design systems in which a fooled model cannot do much harm. Here is what that design looks like.
Regulation and complianceThe EU AI Act and the Cyber Resilience Act for security teams
Two European regulations now define what AI security must look like for products and systems sold in the EU: the AI Act's Article 15, its GPAI systemic-risk duties and its incident reporting, and the Cyber Resilience Act's vulnerability handling and reporting, live since 11 September 2026. Here is how they fit together and what to do.
Agents and tool useSecuring agentic systems: identity, permissions, containment and oversight
Agents inherit whatever access is lying around, act at machine speed and treat instructions as suggestions. This guide sets out the control model that the 2025 and 2026 incidents and the OWASP Agentic Top 10 point to.
Agents and tool useMCP security: how to admit a tool server without handing it the keys
The Model Context Protocol made connecting agents to tools trivial and connecting them safely optional. This guide covers the attack classes, the vulnerability record and an admission process for MCP servers.
Membership is free
Join to propose records and corrections, write for the society and hear when new records and articles are published.