Part 6 — For Inclusive and Diverse Artificial Intelligence This database would enable measurement of the impact of initiatives undertaken to address equality and diversity. Production of these indicators and governance of the database could be led by the CESP (Conseil Supérieur de l’Égalité Professionnelle - Higher Council for professional equality between men and women) within the framework of its task to assess and monitor policies on equality in the workplace. Promoting Transparent Recruitment and Promotion Processes It is crucial to back-up the data policy with a policy for transparent recruitment and promotion processes. Whilst large companies often have such processes in place, this is less common in small businesses and startups. An awareness-raising initiative on transparency in human resources could be implemented in conjunction with FrenchTech. These initiatives could be backed up by the provision of advice to young businesses lacking skills in HR. Carrying out a survey enabling effective measurement of gender discrimination in the digital sector Two female entrepreneurs recently carried out an experiment in which they teamed up with a male business partner, and in doing so observed a significant shift in consideration amongst financers and investors2. This interesting and alarming result deserves to be taken seriously. Assessments could be carried out potentially via A/B testing which would measure the difference between funding obtained with and without a male business partner. All other factors being equal (qualifications, value proposition identity, etc.), the survey could facilitate improved measurement of the existence (or absence) of financial and recruitment biases and indeed any other decisive factors for career progression and the perception of women—by either women or men—in their relationship with ambition, competition and money. Promoting AI research in support of identifying discrimination Solutions capable of identifying workplace discrimination already exist. By analyzing surveys, employee statements, data on salaries and promotion, etc., the company Palantine Analytics was able to detect a number of biases working against women. It would be interesting to promote research which would enable similar solutions to be developed. Setting Aside Funds to Address Diversity Funding for projects working on the development of inclusive and non- discriminatory AI could be established at BPI France or FrenchTech, and equally for digital businesses working on projects with high social and environmental impact 2. Le Monde, “Comment deux entrepreneuses s’inventent un collègue masculin pour convaincre les investisseurs” (How two women invented a male colleague to win over investors), 8 September 2017. 141