Daniela Kluckert is Parliamentary State Secretary at the Federal Minister for Digital Affairs and Transport.
(Photo: Getty Images, Imago)
Technologies for artificial intelligence (AI) are still being developed by teams in companies that are not sufficiently diverse, but predominantly male-dominated. It is not uncommon for serious prejudices to be produced in this way, which can lead to disadvantages for women, among others.
Fortunately, the discriminating AI is a phased-out model. The federal government has set the course for this with its digital strategy: raising awareness, education and diversity are the key to better AI.
According to the World Economic Forum’s Global Gender Gap Report, in 2018 almost 20 percent of women worked in the field of research and development of artificial intelligence worldwide. Since then, the numbers have improved only slightly.
A study by the Deloitte AI Institute speaks of 26 percent of jobs in the field of data and artificial intelligence (AI) that could be filled by women in 2020.
Top jobs of the day
Find the best jobs now and
be notified by email.
In Germany, the situation looks even grimmer. In 2018, just 16 percent of female professionals worked with AI technologies.
Google’s speech recognition recognizes male voices much more often than female ones
Some time ago, Caroline Criado Perez drew attention to the problems that can arise from gender inequality in data sets used by AI systems in her bestseller “Invisible Women” (2019). She refers to a study by the University of Washington, in which scientists found that Google’s speech recognition software was 70 percent more likely to recognize male voices than female ones.
>> Read here: Wake-up call to inventiveness
Apple provided another example in 2019, when the tech company launched its own credit cards. Apple used AI to check who was allowed to receive the credit card. Reports quickly multiplied that men and women were apparently rated differently when the credit card framework was awarded, even though the financial prerequisites were the same.
Prejudices in the AI are also evident in the role assignment. Voice assistance systems, which are mostly female, such as Siri, Alexa and Cortana, are particularly striking.
If algorithms are not only developed by men, but ideally by heterogeneous teams, the chance of recognizing and eliminating discrimination and prejudices increases. Women see things that men don’t, and vice versa. In addition, in diverse teams, the chance of better generalizability and quality of the research results increases.
Girls need to be enthusiastic about research and development as early as daycare
To ensure that more women find their way into AI development and research in Germany, the federal government is promoting educational opportunities in all phases of life, from daycare to university and training and further education.
Well-established programs for schools, such as the little researchers’ house, the “Girls’ Day” and “YouCodeGirls” campaign days, will be continued. The MINT action plan 2 is also intended to expand cooperation, for example between schools and companies, but also to better inform parents so that they can inspire their offspring to do vocational training or study in the MINT subjects (mathematics, computer science, natural sciences and technology).
In short: raising awareness and more education are key factors, on the one hand, to reduce fears and prejudices about new technologies such as artificial intelligence, and on the other hand to get people of different origins, different ages or genders enthusiastic about the subject.
However, education does not only refer to the area of knowledge transfer in schools. Of course, we also need female role models who set a positive example. Initiatives such as “ShetransformsIT” and the “Women in AI” association founded in 2017 are already doing valuable work here.
The discriminating artificial intelligence is an obsolete model
Conclusion: Artificial intelligence can no longer be imagined in our everyday lives and in our future. That’s why it’s so important that we create the right framework conditions so that it’s not just men who enjoy programming algorithms.
We are well on the way to changing that and making the industry more diverse. The discriminating artificial intelligence is an obsolete model.
The author:
Daniela Kluckert is Parliamentary State Secretary at the Federal Minister for Digital Affairs and Transport.
More: Germany is missing around 326,000 mint experts