Risk Classification DatabaseOur open and free Risk Classification Database

Risk classification database

What is our goal?

The appliedAI Institute for Europe gGmbh has the goal to enable trustworthy AI at scale by openly collaborating with all actors of the AI Ecosystem to co-create methods, tools, guidelines, trainings and other resources. Our open and free Risk Classification Database is intended as:

  • empirical reference for anyone seeking to classify an AI System
  • a means of alignment between providers and authorities
  • a data source for academic research on the EU AI Act

Download details and descriptive statistics in this slide deck:

Risk Classification Database September 2024 pdf 1.344 MB

Limitations

  • Changing Classification Criteria: The analysis was done in 2023, hence the “high risk filter” in Article 6(3) was not considered, and reference is made to Annex II, which is now Annex I.
  • Scope of application: The included AI Systems are primarily from the industrial sector.
  • Geographical distribution: The dataset is biased toward Germany and the EU, but some non-EU AI Systems are included too.
  • Limited details: The lack of information about some AI Systems has contributed to an unclear classification.
  • Potential errors: Despite best efforts, the data base may contain erroneous conclusions, or interpretations that may be superseded by more recent work.
  • Automated translations: The content was partially translated using Google Translate and Microsoft Copilot. Thus, the translations may be imperfect, affecting the quality and accuracy of the content.
Title / Business Function Risk Class Transparency Obligation Details

Detection of threats to major events

IT und Sicherheit

inakzeptables Risiko

zutreffend

 Learn more

Biometric / authentication Example 1

Kundenservice

hohes Risiko

zutreffend

 Learn more

Biometric / authentication Example 2

Kundenservice

hohes Risiko

zutreffend

 Learn more

Performance monitoring / observation of employee engagement

Personalwesen

hohes Risiko

zutreffend

 Learn more

Language training for sailors

Kundenservice

geringes Risiko

nicht zutreffend

 Learn more

Medical prognosis

Forschung und Entwicklung

hohes Risiko

nicht geprüft

 Learn more

Natural language processing

Forschung und Entwicklung

geringes Risiko

nicht zutreffend

 Learn more

Medical decision support

Forschung und Entwicklung

hohes Risiko

nicht geprüft

 Learn more

Wind power forecasting

Forschung und Entwicklung

geringes Risiko

nicht zutreffend

 Learn more

Chatbot support for potential candidates

Personalwesen

hohes Risiko

zutreffend

 Learn more