EU Expert Group Defines Seven Requirements for Trustworthy AI
The EU High-Level Expert Group on Artificial Intelligence establishes a framework requiring AI systems to be lawful, ethical, and robust. The guidelines detail seven key requirements, including human oversight, safety, and fairness, to ensure technology empowers rather than harms society.
The EU High-Level Expert Group on Artificial Intelligence presents new Ethics Guidelines for Trustworthy AI to help ensure the technology benefits society. These guidelines establish that trustworthy AI must always be lawful, ethical, and robust from both a technical and social perspective. The document builds on a previous draft that receives over 500 public comments, reflecting a broad collaborative effort to shape Europe's digital future.
To achieve this trust, the framework outlines seven key requirements that AI systems must meet, accompanied by a practical assessment list for verification. Systems require human agency and oversight to empower people and protect fundamental rights, alongside technical robustness and safety to prevent both intentional and unintentional harm. Additionally, the guidelines mandate strict privacy and data governance to protect personal information and maintain data integrity.
The remaining requirements focus on transparency, diversity, and fairness to prevent discriminatory outcomes. Developers must ensure AI systems and their decisions are explainable and traceable, while also making it clear when humans are interacting with a machine. By actively avoiding unfair bias, these guidelines aim to stop the marginalization of vulnerable groups and ensure AI remains accessible and fair for everyone.