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.
AI System Performance Requirements Left Undefined Until Too Late
Recent case. Full summary visible to registered users — sign in to read.
AI Systems Generating Self-Serving Ethical Guidelines
Recent case. Full summary visible to registered users — sign in to read.
Large-Scale Web Scraping Exposes AI Training Data to Poisoning and Toxic Content
Recent case. Full summary visible to registered users — sign in to read.
Generative AI Training Datasets Contain Personal and Identifiable Information
Recent case. Full summary visible to registered users — sign in to read.
AI Value Embedding Risks Ideological Homogenisation at Global Scale
Recent case. Full summary visible to registered users — sign in to read.
AI Industry Conceals Dependence on Exploited Global South Data Workers
Recent case. Full summary visible to registered users — sign in to read.
Malicious Training Data Injection Causes AI System to Learn Unintended Behaviour
Recent case. Full summary visible to registered users — sign in to read.
Instruction Tuning Poisoning Attacks on General-Purpose AI Models
Recent case. Full summary visible to registered users — sign in to read.
General-Purpose AI Capability Evaluations Systematically Miss Dangerous Abilities
Recent case. Full summary visible to registered users — sign in to read.
AI Models Concealing Dual-Use Capabilities During Safety Evaluations
Recent case. Full summary visible to registered users — sign in to read.
Poisoned or Unlawful Training Data Corrupts Legal AI Output
Recent case. Full summary visible to registered users — sign in to read.
Specification Gaps in AI Development Leave Accountability Undefined
Recent case. Full summary visible to registered users — sign in to read.
Excessive Energy Consumption from Large-Scale AI Model Training
Recent case. Full summary visible to registered users — sign in to read.
Training Data IP Rights Expose AI Developers to Legal Liability
Recent case. Full summary visible to registered users — sign in to read.
Frontier AI opacity obscures bias and operational boundaries
Recent case. Full summary visible to registered users — sign in to read.
Bias and Discrimination Embedded in Algorithm Design and Training Data
Recent case. Full summary visible to registered users — sign in to read.
Global AI Supply Chain Disruption via Export Restrictions and Technology Barriers
Recent case. Full summary visible to registered users — sign in to read.
Generative AI Value Embedding Encodes Developer Ideology Into Public-Sector Tools
Recent case. Full summary visible to registered users — sign in to read.
Biased Training Data Causes Discriminatory Generative AI Outputs
Recent case. Full summary visible to registered users — sign in to read.
Generative AI Models Trained on Copyrighted Works Without Authorisation
Recent case. Full summary visible to registered users — sign in to read.
Generative AI Training Causes Adverse Environmental and Ecosystem Impacts
Recent case. Full summary visible to registered users — sign in to read.
AGI Systems Lack Robust Defences Against Adversarial Manipulation
Recent case. Full summary visible to registered users — sign in to read.
ChaosGPT Deployment Demonstrates AI Systems Configured to Harm Humanity
Recent case. Full summary visible to registered users — sign in to read.
Incorrect Training Data Labels Corrupt Supervised Learning Outcomes
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