AIRS-V
Vendor Risk
Evaluate risk signals associated with AI technology providers across their products, customers, and documented activities.
AIRS — AI Incident & Risk ScoreIn development
A clearer picture of AI risk.
Independent, evidence-based risk intelligence for AI vendors, products, and the organisations that deploy them.
The problem
AI risk is difficult to measure. We’re working to change that.
AI is increasingly embedded in hiring, lending, insurance, healthcare, and other consequential decisions.
Yet organisations lack a consistent, independent way to evaluate the risks associated with AI technologies and their real-world use.
Information about incidents, lawsuits, regulatory actions, and deployments is fragmented, making meaningful comparisons difficult.
AIRS is being developed to bring that information together.
Introducing AIRS
Fragmented evidence, one explainable assessment.
The AI Incident & Risk Score (AIRS) is Wayward Fowl’s developing framework for evaluating and comparing AI-related risk using documented evidence and structured intelligence. AIRS is designed to examine multiple risk signals, including:
Incidents
Reported harms, complaints, and adverse outcomes.
Litigation
Lawsuits, allegations, judgments, and settlements.
Regulatory activity
Investigations, enforcement actions, and findings.
Severity
The potential and documented consequences of incidents.
Exposure
The scale and context of AI use.
Transparency
The availability and quality of supporting information.
The goal is to turn fragmented evidence into a more consistent, explainable assessment of AI risk.
Three lenses
Three perspectives on AI risk.
AIRS-P
Product Risk
Examine risk signals associated with specific AI systems, including incidents, litigation, and documented uses.
AIRS-D
Deployer Risk
Understand risk signals associated with organisations using AI in consequential business decisions.
Who it is for
From risk signals to better decisions.
For Law Firms
Identify potential litigation risks, research AI vendors and deployers, and uncover patterns that may warrant further investigation.
For Insurers
Support underwriting research, risk selection, portfolio monitoring, and the development of AI-related insurance products.
For Corporate Risk Teams
Evaluate AI vendors, monitor emerging exposures, and support procurement, compliance, and governance decisions.
For Institutional Researchers
Compare AI technologies and organisations using a consistent framework grounded in documented evidence.
Explainable by design
A score you can investigate.
AIRS is being designed to provide more than a single number.
The framework will aim to distinguish allegations from verified findings, identify the sources behind risk signals, and communicate confidence and data coverage alongside assessments.
- Allegations kept apart from verified findings
- The sources behind every signal
- Confidence and data coverage, stated
A lack of reported incidents does not necessarily mean an AI system is safe.
Our objective is transparent, explainable intelligence — not a black-box rating.
The foundation
Built on WF Intelligence.
AIRS will draw on the connected intelligence Wayward Fowl is building across AI vendors, products, deployments, incidents, litigation, and regulatory activity.
By linking these records to their underlying sources, WF Intelligence is intended to provide the foundation for more informed AI risk assessments.
Early access
Help shape the future of AI risk intelligence.
AIRS is currently under development.
We are seeking input from law firms, insurers, corporate risk professionals, and institutional researchers interested in evaluating the framework and participating in early access.
Request early access
Tell us how you would use AIRS and we’ll come back to you as the framework opens for evaluation.
Development-stage product. AIRS scores and methodology are not yet available for commercial use. Future features and availability are subject to change.