Why is it important to have more women in AI? As part of International Women's Day 2021, Channel Africa interviewed Michelle Beetar, who runs FICO's operations in Africa, about the need for diversity in AI development to fight bias.
As Beetar notes, calls for greater diversity are coming from many groups, including employers. GCHQ, Britain's security and intelligence agency, recently published a white paper on The Ethics of AI. In an editorial for the Financial Times the organization's director, Jeremy Fleming, noted that, "We recognise a team must be inclusive to build good AI."
Diverse teams are more likely to check various ways that data could be biased, leading to AI models that encode that bias. Beetar cites the example of algorithms built to identify what makes a good engineer, at a firm that has a preponderance of male engineers. An AI model trained on that company's data would score males more highly than females.
Beetar notes that FICO has women throughout the analytic development and delivery chain. Last year, Louise Lunn became the company's first leader of global analytics delivery.
As we celebrate International Women's Day today, this interview provides a great example of how AI and analytics can help make for a fairer, more inclusive society.