Labelia, the responsible and trustworthy AI label
Labelia is a complete triptych: framework + assessment + label.
To enable companies to ensure that they are using AI responsibly.
A framework composed of 35+ elements
The evaluation framework for responsible and trustworthy Data Science is developed in open source by many volunteer contributors. It was launched in 2019 by Labelia Labs with the objective of advancing collaborative, responsible, and trustworthy data science. In 2025, the initiative was taken over by datacraft.
A free and open-access self-assessment tool
Building on this framework, a web platform allows any organization to freely conduct a self-assessment to determine its level of maturity regarding responsible and trustworthy AI practices, and to discover resources useful to its activities.
A label for the most advanced companies
An organization may apply for the label once it reaches at least the minimum labeling threshold. datacraft then audits the accuracy and sincerity of its self-assessment. When granted, the label is valid for a period of 2 years. It enables the organization to demonstrate to its stakeholders a high level of maturity and to join a value-driven community committed to responsible and trustworthy AI practices.
Two levels depending on your organization’s maturity
Labelia Intermediate
Labelia Intermediate applies to organizations that achieve a score between 45/100 and 55/100 in the assessment, representing an advanced level of maturity.
Labelia Advanced
Labelia Advanced applies to organizations that achieve a score greater than 55/100 in the assessment. This corresponds to a very advanced level of maturity.
Why obtain the “Labelia – Responsible and Trustworthy AI” label?
Becoming Labelia means…
Anticipating the future European regulation by applying high standards for responsible and trustworthy AI starting today.
Ensuring internally the quality of your practices and discovering new approaches, methods, and tools you may not yet have identified.
Strengthening the trust of your stakeholders: clients, partners, candidates, users, etc.
Who is this label for?
Are you an organization engaged in data science activities?
Do you develop AI models with particular attention to responsibility, trust, and ethical issues, and wish to showcase your efforts in this field?
Becoming Labelia enhances your employer brand among Data Scientists, as well as your reputation with partners and clients, while joining a community of practitioners at the forefront of these topics.
Do you want to evaluate your providers’ practices?
Your providers develop AI models, and you want to ensure that their methods and practices meet the highest standards of responsibility and trust?
Through a high threshold score, verified by a datacraft audit, Labelia-certified companies demonstrate responsible and trustworthy artificial intelligence practices.
Organizations certified “Labelia – Responsible and Trustworthy AI”
Axionable obtained the first “Labelia – Responsible and Trustworthy AI” label, Advanced level, in January 2022.
MAIF is the second company to receive this label awarded by Labelia Labs, also at the Advanced level.
Next came Artefact in spring 2022, at the Intermediate level.
Apricity is the 4th organization to obtain the label, at the Intermediate level, at the end of 2022.
Capgemini Invent (formerly Quantmetry) is the 5th organization to obtain the label, the 3rd at the Advanced level, in March 2023.
Ekimetrics is the 6th organization to obtain the label, the 4th at the Advanced level, in May 2023.
Crédit Agricole Datalab Groupe is the 7th organization to obtain the label, the 5th at the Advanced level, in June 2023.
Gradiant.ai is the 8th organization to obtain the label, the 6th at the Advanced level, and the first internationally, in September 2024.
FLOA is the 9th organization to obtain the label, the 7th at the Advanced level, in January 2025.
AI Builders is the 10th organization to obtain the label, the 8th at the Advanced level, in July 2025.
ContentSide is the 11th organization to obtain the label, the 9th at the Advanced level, in October 2025.
