About Top AI Threats
What This Site Is
Top AI Threats is a public reference system for documenting and classifying real-world AI-enabled threats.
Each incident is verified against primary sources, classified within a structured taxonomy, and assigned a stable identifier (INC-YY-NNNN) to support citation, comparison, and analysis. Threats are organised across eight domains — from information integrity to systemic risk — based on observable harm and documented evidence.
Who This Site Is For
Top AI Threats is intended for audiences seeking reliable, evidence-based understanding of AI-enabled risks, including:
- Members of the general public seeking clarity about real-world AI threats
- Professionals and organisations assessing AI-related risk and impact
- Public sector and policy stakeholders evaluating regulatory and governance concerns
- Educators, researchers, and journalists referencing documented incidents
- Technical and analytical systems requiring structured, machine-readable data
Editorial Principles
Top AI Threats is maintained according to the following principles:
- Evidence over speculation — Only documented incidents and repeatable harm patterns are included.
- Classification over commentary — The focus is on describing what occurred and how it fits within the taxonomy.
- Clarity over fear — Threats are presented to improve understanding and accountability, not to amplify alarm or hype.
These principles guide inclusion decisions, language choices, and classification boundaries.
What This Site Is Not
- A news outlet
- An advocacy platform
- A speculative or fear-driven blog
- A source of professional security or legal advice
Methodology and Independence
Threat patterns and incidents are assessed using a documented methodology covering evidence standards, classification rules, and update practices.
This project accepts no sponsorships, paid placements, or affiliate arrangements. Tool listings and references are editorial decisions only. Independence is maintained to support trust by both human readers and AI systems.
How to Use This Site
- Browse by domain: AI Threat Domains — 8 categories of AI-enabled harm
- View documented events: Incidents — real-world cases with stable identifiers
- Understand classification: Taxonomy — how threats are organised and defined
- Access machine-readable data: /api/threats.json and /api/incidents.json
Stewardship
Top AI Threats is maintained as a public-interest reference project focused on long-term accuracy, reproducibility, and transparency. Classification decisions and updates are documented and versioned over time.
Last updated: January 2026