Over the past few years, the AI governance landscape in Europe has advanced substantially. With the publication of the AI Act in the official journal in July 2024, the European Union has made a significant move to regulate and govern the use of artificial intelligence (AI) systems. Standards that offer a presumption of conformity—particularly for high-risk systems—are expected to progressively emerge.

Improving practices

Companies and industries are seeking concrete solutions to improve AI practices. The objective is both to enhance competitiveness and align with the AI Act, including for non-high-risk systems. Evaluation frameworks with clear indicators and clear scores are required for various topics such as Human agency and oversight / Technical robustness and safety / Privacy and data governance / Transparency / Diversity, non-discrimination, and fairness / Social and environmental well-being.

Different AI labels based on different evaluation frameworks can now be found. Some are product-oriented, others are organization-focused, while others cater to specific industry sectors.

Joining forces

To promote a unified European approach, Confiance.ai has joined forces with key European-oriented initiatives. In 2022, a collaboration was initiated with the German institute VDE, the French association Positive AI, and IEEE. The common aim is to avoid fragmentation and promote a common European method to evaluate AI with clear labels.

A strong alliance

A joint specification gathering the indicators from the VDESPEC 90012, the Positive AI framework and IEEE’s CertifAIED framework has been successfully developed. An AI Trust Alliance is constructed to continue this work and offer strong labels for AI products, organizations, and individuals.

The joint specification is intended to be straightforward to implement, to complement harmonized standards, to remain open to everyone (associations, companies), and to be recognized by public institutions (particularly European).

Benefits

The alliance will benefit:

• companies to communicate on product characteristics and target market segments, to qualify suppliers, to improve internal processes
• institutions to limit fragmentation, to offer “codes of conduct”
• and to other labeling organizations (to join forces, to easily develop new services).

The alliance highlights the need for concrete technological solutions such as the methods and tools developed by Confiance.ai.

By working together, Europe is paving the way for responsible AI governance that not only complies with regulatory frameworks but also promotes innovation and competitiveness.

For more information, or to get involved in this collaborative effort, please contact us at contact@confiance.ai.