EU-AIA-15-06 Requirement 15 — Accuracy, robustness and cybersecurity

Address feedback loops with appropriate mitigation measures

Description

High-risk AI systems that continue to learn must ensure that any feedback loops are duly addressed with appropriate mitigation measures.

Sign in to view the full obligation text, AI-generated applicability analysis, evidence checklists, and compliance mapping.

Design and develop for appropriate accuracy, robustness and cybersecurity

Declare accuracy levels and metrics in instructions

Ensure resilience to errors, faults and inconsistencies

Implement technical and organisational resilience measures

Eliminate or reduce biased feedback loops in learning systems

Ensure resilience against unauthorized third-party alteration

Implement appropriate cybersecurity technical solutions

Address AI-specific vulnerabilities with technical solutions

ReguLume automatically maps regulatory obligations to your system inventory, identifies compliance gaps, and generates remediation plans.

Get Started