1. Context and Functionality
Klara AI conducts automated interviews with applicants and processes their responses. The information collected includes, among others, salary expectations, teamwork skills, motivation, and other data relevant to the recruiting process. The responses are documented and stored in the recruiter system for further processing.
2. Risk Classification
According to the EU AI Act, AI systems fall into four categories:
-
Unacceptable risk – prohibited
-
High risk – subject to strict regulation
-
Limited risk – transparency obligations
-
Minimal or no risk – largely unregulated
Klara AI, as a chatbot for candidate interviews, clearly falls under the category “Limited risk.” Chatbots are subject to transparency requirements, meaning users must be informed that they are interacting with an AI system.
3. Required Measures & Compliance Obligations
3.1 Transparency
-
Disclosure: Klara AI must clearly indicate that it is an AI system (e.g., via a message at the beginning of the interview).
-
Documentation: Provide clear information about the system’s purpose, functions, and processes.
3.2 Data Quality and Governance
-
Ensure training data and interview logic are relevant, representative, and as error-free as possible.
-
Maintain traceability regarding how and where data is stored (e.g., recruiter system).
3.3 Human Oversight
-
Klara AI must only support decision-making. Human review and intervention must remain possible at all times (e.g., during candidate evaluation).
3.4 Security & Robustness
-
Guarantee technical safeguards to protect against misuse or data breaches.
-
Ensure reliable operation under load and secure storage of applicant data.
4. Conclusion & Recommendations
| Risk Category | Assessment for Klara AI |
|---|---|
| Category | Limited Risk |
| Legality | Permitted, subject to transparency and documentation obligations |
| Obligations | Transparency notice, documentation, data quality, human oversight, and technical safeguards |
Summary: Klara AI does not qualify as a “High-Risk” system under the EU AI Act. Therefore, extensive conformity assessments are not required. However, compliance with transparency obligations, documentation standards, and security measures is mandatory.