In order to provide committed actors with a practical tool to design, develop and manage artificial intelligence systems that respect fundamental human rights, Numeum, together with its partners, is launching the Ethical AI project. This initiative is based on three major actions: a manifesto to formalize the commitment of the players, a practical guide to rely on a methodology and tools, and a community of members who will have the opportunity to exchange ideas to progress.

A project led by Numeum for ethical and responsible artificial intelligence

At the end of 2020, six collective intelligence sessions were conducted and brought together 360 participants (coders, developers, data scientists, data architects, data engineers, AI product managers, testers, DPOs or lawyers, etc.). They all participated in workshops to dissect the main theoretical principles of ethical AI and to imagine ways to apply them in practice to the various components of the technology (algorithm, data, etc.). These six exchanges resulted in a Practical Guide that proposes a useful method for operationalizing the major ethical principles during the design, development and deployment of AI solutions.

It is this guide that is the anchor of the Ethical AI initiative. This project, led by Numeum, brings together a network of partners (professional organizations, schools, interdisciplinary AI institutes, associations, foundations, think tanks, etc.) committed to a more responsible digital world. Its ambition is to accompany AI experts who wish to integrate ethics into the heart of their activities in order to support the operational implementation of general ethical principles, based on existing work, into practical methods and tools.

Supported by the Coordinator of the National Strategy for AI, Renaud VEDEL, Numeum’s initiative enables to federate a network of actors from different AI universes. Several partners are also contributing: the 3IA Côté d’Azur institute, Telecom Valley, Impact IA, Hub France IA, the Grand-Est region, Grand E-Nov+, Institut Dataia, ANITI, Aivancity, the 42 school and the French government.

A project based on a manifesto and a practical guide for ethical AI

By committing to the Ethical AI manifesto, signatories have free access to the Practical Guide for Ethical AI and join the community of actors committed to the development and promotion of more responsible AI solutions. Companies, organizations of the digital ecosystem or actors of training in AI systems, all can formalize their commitment by signing the Manifesto.

The signatories of the Manifesto for Ethical AI commit to:

• Implement the methodology and, insofar as they are applicable, the recommendations presented in the Practical Guide to Ethical AI published by Numeum and co-constructed with representatives of the AI ecosystem in France.
• To promote the ethical dimension of the AI solutions that the organization designs and/or deploys and/or administers and/or trains.
• Contribute to the improvement of the state of the art in terms of ethics in AI, in particular by the participation of a representative of the organization in the revision and enrichment of the guide. This work will be based on the elements shared by the members of the collective of signatories of the Manifesto – e.g.: feedback, theoretical advances, etc.

Around thirty associative, academic or professional actors have already joined the project: Accenture, Aivancity, AIvidence, Alter Way, Astek (group), CREEIA, DATAVALORIS, DEVOTEAM, Docaposte, EITA Consulting, Horizon Software, Keyrus, La Clinique De L’Ethique (Factonics), LINAGORA, Linedata, Myriad, Orange Business Services, Quanmetry, SAASWEDO, Sage, Seenapsys, TAAS, Talkr.ai by Kwalys, The Neoshields, The Voice Lab, Utocat, Ysance.

Translated from Lancement de l’initiative “Ethical AI” pour une IA éthique, responsable et digne de confiance