AI Risk Assessment Tool

This project evaluates the risk of an AI system using principles from the AIRO Ontology. It supports command-line and web-based (Sinatra) interfaces and outputs results in both Markdown and RDF/Turtle formats.

---

Features

• Collects AI risk inputs via a structured questionnaire.
• Applies heuristic risk scoring logic.
• Outputs:
  • Human-readable Markdown Report
  • RDF/Turtle data for SPARQL/linked data applications
• Available as a CLI tool or web app (Sinatra).

---

Installation

git clone https://github.com/inloopstudio/AI-Risk-Ontology-AIRO.git
cd AI-Risk-Ontology-AIRO
bundle install

---

Usage

1. Command-Line

Generate questionnaire:

 bundle exec ruby ai_risk_assessment.rb generate_questionnaire > questionnaire.txt

Fill questionnaire.txt, then:

 bundle exec ruby ai_risk_assessment.rb process_responses questionnaire.txt

Outputs:

• assessment_report.md
• assessment_report.ttl

---

2. Web Application

Start the Sinatra server:

bundle exec ruby app.rb

Visit http://localhost:4567, fill out the form, and get your AI Risk Report in Markdown.

---

3. Replit Deployment

The app is avaible to remix at https://replit.com/@inloop/AI-Risk-Ontology-AIRO

---

Output Formats

Markdown

# AI Risk Assessment Report
...
**Total Risk Score:** 13  
**Overall Risk Level:** HIGH

RDF/Turtle

@prefix airo: <http://example.org/airo#> .

<http://example.org/ai_system/1>
  a airo:AISystem ;
  airo:hasDomain "Law Enforcement" ;
  ...
  airo:hasRiskLevel "HIGH" .

---

Sponsors

This project is sponsored by inloop.studio

License

MIT