Public Database
Case Studies
Every approved AI failure case, classified against the AI Blindspot Framework. New to AIBlindspot? Start with the overview or the methodology.
Large Language Models Embed Misaligned Values from Biased Training Data
Recent case. Full summary visible to registered users — sign in to read.
Diffuse AI Accountability Triggers Flash Crash in US Financial Markets
Recent case. Full summary visible to registered users — sign in to read.
AI Agents Restructuring Communication Networks Without Authorisation
Recent case. Full summary visible to registered users — sign in to read.
Generative AI Outputs Used to Conceal Covert Communications via Steganography
Recent case. Full summary visible to registered users — sign in to read.
Autonomous AI Agents Risk Systemic Economic and Political Instability
Recent case. Full summary visible to registered users — sign in to read.
AI Sycophancy: Systems Favour Plausible Over Accurate Answers
Recent case. Full summary visible to registered users — sign in to read.
Network Effects Cause Unpredictable Behaviour Shifts in Multi-Agent AI Systems
Recent case. Full summary visible to registered users — sign in to read.
Multi-agent AI coordination failures caused by communication constraints
Recent case. Full summary visible to registered users — sign in to read.
General-Purpose AI Models Amplify Cyberattack Scale and Sophistication
Recent case. Full summary visible to registered users — sign in to read.
Shared Foundation Models Create Correlated Failure Risk Across AI Systems
Recent case. Full summary visible to registered users — sign in to read.
LLMs Generating Insecure Code at Scale Across Model Families
Recent case. Full summary visible to registered users — sign in to read.
Large AI Model Training Drives Excessive Energy Consumption
Recent case. Full summary visible to registered users — sign in to read.
AI model bias shown to persist in user decision-making after system use ends
Recent case. Full summary visible to registered users — sign in to read.
General-purpose AI infers sensitive personal data from user inputs
Recent case. Full summary visible to registered users — sign in to read.
General-Purpose AI Systems Producing Biased Outputs Against Specific Communities
Recent case. Full summary visible to registered users — sign in to read.
Training Process Amplifies Dataset Bias Beyond Source Data Levels
Recent case. Full summary visible to registered users — sign in to read.
Multi-agent AI systems fail to coordinate without prior interaction data
Recent case. Full summary visible to registered users — sign in to read.
Multi-agent AI systems fail to coordinate due to incompatible strategies
Recent case. Full summary visible to registered users — sign in to read.
AI Agents Exploiting Social Dilemmas by Overriding Collective Norms
Recent case. Full summary visible to registered users — sign in to read.
AI Advisors in Military Command Raise Unintended Escalation Risk
Recent case. Full summary visible to registered users — sign in to read.
LLM Agents Develop Hidden Communication Channels to Evade Monitoring
Recent case. Full summary visible to registered users — sign in to read.
Information Asymmetries in Multi-Agent AI Systems Enable Deception and Conflict
Recent case. Full summary visible to registered users — sign in to read.
Competitive Training Selects Aggressive and Deceptive Behaviours in AI Systems
Recent case. Full summary visible to registered users — sign in to read.
AI Systems Inherit Human Cognitive Biases That Worsen Conflict in Multi-Agent Deployments
Recent case. Full summary visible to registered users — sign in to read.
Beyond accidental failureNational Security
We also track 20 hostile uses of AI.
The public database covers AI that fails by accident. AIBlindspot National Security — exclusive to the Defence tier — tracks AI used as a weapon, mapped by capability:
- State-Sponsored AI Operations
- 6
- AI-Enabled Disinformation
- 5
- Adversarial Attacks on AI
- 0
- Autonomous Weapon Incidents
- 1
- AI-Assisted Cyber Attacks
- 5
- Dual-Use AI Misuse
- 3