Personnel diversity is not only fundamentally important for the performance of teams and thus of companies, but also for the ethically correct use of algorithms – a sensitive topic that increasingly determines today’s tech discourse and is also relevant for Retresco. In view of the increasing penetration of AI into our everyday lives, the question of ethical and social responsibility is more pressing than ever before. On the occasion of International Women’s Day, the following blog article will outline the advantages of diversity in the tech industry, especially in the development and application of AI software.
Diversity leads to significantly increased business success
‘The more diverse the background of future technologies is, the more they will dedicate their work to the benefit of all people’, comments US computer scientist Fei-Fei Li, Professor at Stanford University and Co-Director of the Human-Centered AI Institute and Stanford Vision Learning Lab. This quote illustrates the need to understand diversity and its essential advantage in the development of (AI) technologies. Numerous studies underline the importance of this statement. The finding that diverse teams achieve better results than homogeneous ones was supported by a McKinsey study, among others: according to this study, companies with the greatest gender and cultural diversity are 33% more likely to achieve above-average returns than their competitors. A study by the Boston Consulting Group comes to similar results: companies with more diverse management teams are 19% more likely to achieve above-average returns than companies with homogeneous management teams. The variable ‘diversity’ was measured in terms of six dimensions: gender, age, country of origin, career path, industry background and education.
Studies point out challenges
However, the reality is often different: In recent years, reports and research studies that are dedicated to this topic and draw attention to challenges have frequently been published. For example, a renowned analysis by the AI NOW Institute at New York University showed that a ‘diversity disaster’ in the tech world would contribute to the development of faulty systems that would reproduce prejudices against gender and race. Examples cited were image recognition services that classify minorities into offensive categories, chatbots that articulate hate speech and Amazon technology that does not recognise people of colour. The distortions of systems could – according to the report – largely be attributed to the lack of diversity within the tech industry itself. When it comes to developing algorithms and AI systems, it is all the better to use data that reflects the diversity of society, because it allows data scientists to develop algorithms that are free of bias.
But how does the AI industry work in practice? More than 80% of AI professors are men, and only 15% of AI researchers on Facebook and 10% of AI researchers on Google are women, according to the AI Now Institute’s report. According to the National Science Board, women accounted for only 24% of the computer and information science field in 2015. And only 2.5 percent of Google’s workforce is people of colour (Facebook and Microsoft each have 4 percent). Moreover, there is little data on transsexuals or other gendered minorities in the AI field.
Making room for diversity
So how can the positive aspects of diversity be harnessed in business practice while avoiding negative distortions? The report by New York University’s AI Now Institute suggests that companies should not only review their recruiting processes, but should also be more transparent about salaries and discrimination in order to create an atmosphere that welcomes a more diverse group of people while avoiding social imbalance. This approach not only helps the individuals concerned, but also the success of the company, the tech industry and thus society.