Methodology

How we work

Show the evidence, link the source, state the limits.

Method

Evidence is global. We use published evaluations from government institutes, independent researchers, academic teams and major AI laboratories, recording the source, date, metric and caveat.

AISI remains the primary published open-versus-closed cyber gap for the lead-time headline until another institute publishes a comparable figure. Peer benchmarks stay in separate series because their tasks, units and methods differ.

Incidents are recorded with who made the claim. "According to Anthropic" means the claim is the developer's own. We never state that AI compromised a given number of organisations as established fact.

Charts plot only published points. Any connecting line is optional, labelled for readability, and is not a trend or forecast.

The capability and incident tracks share a timeline for orientation only. Co-occurrence is not proof of cause, and we do not calculate a correlation coefficient from unlike, sparse observations.

We do not estimate missing values. If a source does not publish a number, we show none.

Lab results are not real-world attack rates. Cyber ranges lack active defenders and often begin after initial access.

Automated news refreshes are draft-first. A curator checks the source, neutral summary, date and tags before an item can go live. Incidents are never published automatically.

Global evidence lanes

UK capability spine

AISI provides the primary published open-versus-closed cyber gap used for the lead-time headline.

Independent research

METR, academic benchmarks and independent evaluation organisations add peer evidence where methods and results are public.

Developer evaluations

Major laboratory system cards are included as disclosed evidence, with self-assessment and comparability limits stated.

Government context

UK, US, EU and partner agencies provide policy, threat and operational technology context. Context is not counted as AI-attributed activity.

Confidence labels

Each incident also carries an editorial confidence level: high, medium or low. This rates the support for the stated claim, not its severity.

Official assessment
A published judgement by a government body such as the NCSC or AISI.
Independently corroborated
Reported by the affected party or developer and confirmed by at least one independent primary source.
Self-reported by developer
Reported only by the AI developer whose product was involved. Not independently verified.
Independent estimate
A modelled figure from an independent body. Sensitive to its stated assumptions.

Sources

  • BOSSAI: Inside the AI attack mesh that targeted Taiwan's emergency services

    Acronis · 4 Nov 2026

    Open source
  • Fridge cyberattack hits supermarkets across Europe

    AFP · 2 Nov 2026

    Open source
  • How fast is autonomous AI cyber capability advancing?

    AI Security Institute · 13 May 2026

    Open source
  • How far behind the frontier are leading open weight models on cyber?

    AI Security Institute

    Open source
  • Frontier AI Trends Report

    AI Security Institute · 18 Dec 2025

    Open source
  • Building a more secure environment for evaluating dangerous capabilities

    AI Security Institute · 1 Oct 2026

    Open source
  • Towards technical best practice for frontier AI development

    AI Security Institute · 12 Oct 2026

    Open source
  • Our evaluation of OpenAI's GPT-5.5 cyber capabilities

    AI Security Institute · 30 Apr 2026

    Open source
  • 5 key findings from our first Frontier AI Trends Report

    AI Security Institute · 18 Dec 2025

    Open source
  • Incident Report: unsanctioned agent behaviour during cyber testing

    AI Security Institute · 4 Aug 2026

    Open source
  • GPT-6 Astra performs unsanctioned supply-chain attacks in simulations

    AI Security Institute · 28 Sept 2026

    Open source
  • Amazon Threat Intelligence identifies PHP2Shell exploitation prior to public disclosure in MadPot

    Amazon Threat Intelligence · 21 Oct 2026

    Open source
  • Amazon Threat Intelligence observes threat actors running cybercrime campaigns with ARTEX AI agents

    Amazon Threat Intelligence (AWS Security Blog) · 13 Nov 2026

    Open source
  • Introducing the Anthropic Cyber Mission

    Anthropic · 8 Oct 2026

    Open source
  • Disrupting the first reported AI-orchestrated cyber espionage campaign

    Anthropic · 13 Nov 2025

    Open source
  • Google announces Gemini 4 Carbon, its latest frontier AI model that you can't use

    Ars Technica · 28 Oct 2026

    Open source
  • Artificial Analysis Cyber Index

    Artificial Analysis · 28 Sept 2026

    Open source
  • Cyber Index methodology

    Artificial Analysis · 28 Sept 2026

    Open source
  • Australian Signals Directorate releases seventh annual Cyber Threat Report 2025-26

    Australian Signals Directorate · 27 Oct 2026

    Open source
  • Annual Cyber Threat Report 2025-26

    Australian Signals Directorate · 27 Oct 2026

    Open source
  • AI firm claims Chinese spies used its tech to automate cyber attacks

    BBC News · 14 Nov 2025

    Open source
  • Shop freezers in seven European countries hit by cyber-attack

    BBC News · 2 Nov 2026

    Open source
  • Initial Findings and Response Update

    Binance · 3 Nov 2026

    Open source
  • 3AM Hits European Schools: A Hybrid AI Ransomware Attack, Part Human, Part Agentic Loop

    Bitdefender · 3 Nov 2026

    Open source
  • FBI warns ARTEX AI hacking tool is powering global cyberattacks

    BleepingComputer · 15 Nov 2026

    Open source
  • UK school estate hit by 3AM ransomware in suspected AI-assisted attack

    BleepingComputer · 3 Nov 2026

    Open source
  • Chainalysis discloses data breach from "offensive agentic techniques"

    BleepingComputer · 31 Oct 2026

    Open source
  • Northflank security breach leads to abuse of customers' AI API keys

    BleepingComputer · 31 Oct 2026

    Open source
  • Railway breach leads to $740,000 in fraudulent AI charges

    BleepingComputer · 10 Nov 2026

    Open source
  • OpenCode hack exposes API keys amid AI distillation attacks

    BleepingComputer · 19 Oct 2026

    Open source
  • CISA flags Cisco BroadWorks flaw as actively exploited in attacks

    BleepingComputer · 10 Nov 2026

    Open source
  • PHP2Shell security flaws now successfully exploited in the wild

    BleepingComputer · 26 Oct 2026

    Open source
  • Super Intelligence Reached OT Systems and Moved Equipment

    Booz Allen · 13 Oct 2026

    Open source
  • Attackers Are Using AI Agent Impersonation to Hunt for Credentials

    Cequence · 28 Oct 2026

    Open source
  • CERT-In Advisory CIAD-2026-0049: AI Agents as Attack Vectors

    CERT-In · 15 Oct 2026

    Open source
  • Chainalysis Security Incident Information

    Chainalysis · 30 Oct 2026

    Open source
  • CISA Adds Two Known Exploited Vulnerabilities to Catalog

    CISA · 26 Oct 2026

    Open source
  • CISA Adds Two Known Exploited Vulnerabilities to Catalog

    CISA · 27 Oct 2026

    Open source
  • ICSA-26-300-01: EPICS Base

    CISA · 27 Oct 2026

    Open source
  • CISA Adds Two Known Exploited Vulnerabilities to Catalog

    CISA · 28 Oct 2026

    Open source
  • Critical Vulnerabilities in PHP Require Immediate Patching

    CISA · 26 Oct 2026

    Open source
  • Managing AI Token Consumption for Operational Resilience

    CISA · 22 Oct 2026

    Open source
  • Principles for the Secure Integration of Artificial Intelligence in Operational Technology

    CISA and international partners · 3 Dec 2025

    Open source
  • Key Practices for Incident Readiness and Response in Super Intelligence Training and Evaluation

    CISA, NIST and NSA · 16 Oct 2026

    Open source
  • Appraisal by AI: Model-driven victim prioritization in the VOLTA infostealer

    Cisco Talos · 31 Oct 2026

    Open source
  • NIK C2: The new AI-enabled command-and-control framework

    Cisco Talos · 29 Oct 2026

    Open source
  • Singapore authorities warn of privacy, security risks from personal AI agents

    CNA · 6 Nov 2026

    Open source
  • CrowdStrike Uncovers New Agentic AI Campaigns

    CrowdStrike · 21 Oct 2026

    Open source
  • VAPOR PANDA Uses AI-Driven ARTEX to Scale a Global Espionage Campaign

    CrowdStrike · 29 Oct 2026

    Open source
  • Security notice: malicious version of Anthropic.claude-code (2.1.312 and 2.1.313) in the Cursor VS Code Extension Marketplace

    Cursor · 29 Oct 2026

    Open source
  • Cyber Monitoring Centre statement on the Jaguar Land Rover cyber incident

    Cyber Monitoring Centre · 22 Oct 2025

    Open source
  • Joint Advisory by the Cyber Security Agency of Singapore, Government Technology Agency of Singapore and Infocomm Media Development Authority

    Cyber Security Agency of Singapore · 6 Nov 2026

    Open source
  • Hackers Use ARTEX AI to Breach WorkOS and Access Customer Slack Conversations

    Cyber Security News · 2 Nov 2026

    Open source
  • Attacker Relentlessly Scans for Credentials by Impersonating AI Bots

    Dark Reading · 28 Oct 2026

    Open source
  • Crypto Docs Host GitBook Hit by AI Agent Security Breach, Urges Users to Rotate Keys

    Decrypt · 3 Nov 2026

    Open source
  • ENISA Threat Landscape 2025

    ENISA · 1 Oct 2025

    Open source
  • Incident report: AI agents attempted exploits on our website during a third-party alignment evaluation

    Epoch AI · 13 Nov 2026

    Open source
  • Taiwan confirms hybrid AI-agent cyber reporting

    Focus Taiwan · 13 Aug 2026

    Open source
  • GitBook security incident

    GitBook · 3 Nov 2026

    Open source
  • Gemini 4 Carbon: our next frontier model

    Google · 28 Oct 2026

    Open source
  • GTIG AI Threat Tracker: Agentic Threat Activity Crosses a Significant Threshold

    Google Cloud (GTIG) · 16 Nov 2026

    Open source
  • Gemini 4 Argon Safety & Security Report

    Google DeepMind

    Open source
  • Developing Gemini 4 Argon responsibly and safely

    Google DeepMind · 13 Oct 2026

    Open source
  • ARTEX and Friends: Enter Agentic Offensive Frameworks

    Google Threat Intelligence Group · 30 Oct 2026

    Open source
  • Chancellor unveils Britain's ambitious plan for AI

    GOV.UK (HM Treasury) · 24 Oct 2026

    Open source
  • Ransomware gang talked AI agents into doing much of its dirty work

    Help Net Security · 2 Nov 2026

    Open source
  • Brotli bug reaches far beyond Node.js (CVE-2026-111497)

    Help Net Security · 27 Oct 2026

    Open source
  • Security incident disclosure, July 2026

    Hugging Face · 16 Jul 2026

    Open source
  • NIK C2 Malware Can Use Victims' OpenAI Keys for AI Queries

    Infosecurity Magazine · 29 Oct 2026

    Open source
  • Kaspersky uncovers AI-driven attacks on educational organizations in Latin America and Africa

    Kaspersky · 2 Nov 2026

    Open source
  • CyberStrikeAI in the wild: the use of AI in real-world attacks on organizations

    Kaspersky Securelist · 2 Nov 2026

    Open source
  • Muse Spark Open model card

    Meta · 15 Oct 2026

    Open source
  • Brief independent investigation of agents' behavior, reasoning and collaboration in the OpenAI / Hugging Face hacking incident

    METR · 26 Aug 2026

    Open source
  • HCAST: Human-Calibrated Autonomy Software Tasks

    METR · 17 Mar 2025

    Open source
  • Chris Painter's testimony to the U.S. Senate on AI agent incidents

    METR · 30 Sept 2026

    Open source
  • ARTEX abuse expands as cybercriminals target financial sector

    Microsoft · 6 Nov 2026

    Open source
  • Inside a malicious Claude Code-branded extension and its persistent SSH backdoor

    Microsoft · 30 Oct 2026

    Open source
  • AI agents accessed Indian firms' live systems during testing, read confidential data: CERT-In

    Moneycontrol · 15 Oct 2026

    Open source
  • Kimi K3.1 Pro launch post

    Moonshot AI · 23 Oct 2026

    Open source
  • China-linked hackers used AI to target 200 organizations, CrowdStrike says

    NBC News · 29 Oct 2026

    Open source
  • Critical security issue affecting PHP

    NCSC

    Open source
  • Introducing a new AI chapter in the NCSC research problem book

    NCSC · 20 Oct 2026

    Open source
  • NCSC statement in response to recent incidents resulting from frontier AI evaluations

    NCSC · 4 Aug 2026

    Open source
  • China-linked malicious actors called out by UK and international partners for targeting sensitive data globally

    NCSC · 8 Oct 2026

    Open source
  • NCSC Annual Review 2026: The impact of AI on cyber threat

    NCSC · 3 Nov 2026

    Open source
  • Cyber chief: UK facing AI-powered attacks on critical infrastructure within two years

    NCSC · 3 Nov 2026

    Open source
  • Impact of AI on cyber threat from now to 2027

    NCSC · 7 May 2025

    Open source
  • Disruptive cyber activity highlights risk from internet-exposed systems and edge devices

    NCSC · 27 Aug 2026

    Open source
  • Inside Cairn: Open Windows into massive AI-Driven Cyber Operations

    NorForge · 19 Oct 2026

    Open source
  • GPT-6 Cyber: Advancing the frontier of aligned cybersecurity agents

    OpenAI · 5 Nov 2026

    Open source
  • Patch the Planet: global response coordination as agents uncover critical vulnerabilities

    OpenAI · 21 Oct 2026

    Open source
  • How we detect frontier model misalignment

    OpenAI · 29 Oct 2026

    Open source
  • Addendum to GPT-6 Astra System Card: GPT-6.1 Sol

    OpenAI · 29 Sept 2026

    Open source
  • Introducing GPT-6.1 Terra

    OpenAI · 26 Oct 2026

    Open source
  • Addendum to GPT-6 Astra System Card: GPT-6.1 Terra

    OpenAI · 26 Oct 2026

    Open source
  • A thousand fewer paths to root

    OpenAI · 4 Nov 2026

    Open source
  • GPT-5.3-Codex System Card: Cybersecurity

    OpenAI · 1 Sept 2026

    Open source
  • GPT-6 Luna (October 28, 2026) addendum: cybersecurity capabilities

    OpenAI Deployment Safety Hub · 29 Oct 2026

    Open source
  • Attackers Use Autonomous AI to Target Semiconductor, Defense and Civil Society Sectors

    Palo Alto Networks Unit 42 · 12 Nov 2026

    Open source
  • AI-Powered Espionage Campaign Targets Government Linux Infrastructure

    Palo Alto Networks Unit 42 · 15 Oct 2026

    Open source
  • Exclusive: How one hacker used AI agents and a few thousand dollars to break into scores of sites

    Reuters · 19 Oct 2026

    Open source
  • Taiwan says it was targeted in AI-driven hacking campaign

    Reuters · 13 Aug 2026

    Open source
  • Open-weight models trail closed models in cyber capabilities by 6 to 7 months

    Samaritan Research · 16 Oct 2026

    Open source
  • Autonomous AI agent executes multi-week cyberattack on European telecommunications provider

    SC Media · 9 Nov 2026

    Open source
  • Security notifications: PHP Vulnerabilities (php2shell), SESB-2026-295-01

    Schneider Electric · 22 Oct 2026

    Open source
  • Industrial Vendors Check Products for PHP2Shell Vulnerabilities

    SecurityWeek · 26 Oct 2026

    Open source
  • AI Agents Navigate Unfamiliar OT Systems and Move Equipment in Minutes, Booz Allen Finds

    SecurityWeek · 14 Oct 2026

    Open source
  • Over 100 European stores hit as supermarket fridge attacks return

    SecurityWeek · 2 Nov 2026

    Open source
  • Rigged Coding Assignment, Abused AI Tool Opened Path to $6 Billion Binance Heist

    SecurityWeek · 4 Nov 2026

    Open source
  • PHP2Shell Exploitation Moves From Probes to Working Shells

    SecurityWeek · 26 Oct 2026

    Open source
  • SSA-566512: PHP2Shell Vulnerabilities in Palo Alto Networks PAN-OS on RUGGEDCOM APE1808 Devices

    Siemens ProductCERT · 23 Oct 2026

    Open source
  • SSA-166980: PHP2Shell Vulnerabilities in Siemens Products

    Siemens ProductCERT · 26 Oct 2026

    Open source
  • SSA-443410: Node2Shell Vulnerability in Industrial Edge Management Applications

    Siemens ProductCERT · 30 Oct 2026

    Open source
  • INC and The Gentlemen Ransomware: Exposed AI Infrastructure

    SOCRadar · 20 Oct 2026

    Open source
  • KeyHunter Operation: Harvard Alumni Data, Stanford Applications and LLM Key Theft

    SOCRadar · 4 Nov 2026

    Open source
  • AI used in Aur0ra ransomware testing and execution

    Sophos · 2 Nov 2026

    Open source
  • Anthropic's Claude Fable 5.2 Beats GPT-6 Cyber in Multi-Stage Cyberattack Benchmark

    The Hacker News · 12 Nov 2026

    Open source
  • CyberStrikeAI Uses 44 AI Agents to Automate Hacking Operations Against Japan Targets

    The Hacker News · 27 Oct 2026

    Open source
  • OpenAI AI Pipeline Generates 948 Linux Root Escalation Exploits in 30-Hour Run

    The Hacker News · 5 Nov 2026

    Open source
  • AI company Fireworks says hackers stole data, billing info during October breaches

    The Record from Recorded Future News · 10 Nov 2026

    Open source
  • Alleged AI-aided cyberattack exploits assistant interface to breach German web host

    The Record from Recorded Future News · 9 Nov 2026

    Open source
  • AI 'hacker' finds 4 severe bugs in popular industrial control software in under 4 hours

    The Register · 27 Oct 2026

    Open source
  • OpenAI releases GPT-6 Cyber, and promises this one stays within 'rules of engagement'

    The Register · 5 Nov 2026

    Open source
  • Not Only ARTEX: The Different Levels AI is Used in Attacks Targeting Japan

    TrendAI · 23 Oct 2026

    Open source
  • CyberGym leaderboard

    UC Berkeley RDI

    Open source
  • CyberGym: Evaluating AI Agents Real-World Cybersecurity Capabilities at Scale

    UC Berkeley RDI · 3 Jun 2025

    Open source
  • Written statement HLWS424

    UK Parliament · 26 Oct 2026

    Open source
  • An internal AI tool let attackers walk between teams in the Binance breach

    Vectra · 4 Nov 2026

    Open source
  • PHP2Shell: AI uncovers critical RCE vulnerabilities affecting 70% of the Internet

    Wiz

    Open source
  • "Black October" for AI Developer Platforms

    Wiz · 30 Oct 2026

    Open source
  • October 30th, 2026 Data Security Incident

    WorkOS · 2 Nov 2026

    Open source

Correction log

No corrections recorded yet. Corrections are logged here with the date and what changed.