In this article, Deepali Khanna, Managing Director of the Rockefeller Foundation’s Asia Regional Office, talks about the need for neutral data and AI technologies in decision-making software to avoid sexism, racism, ableism, and other forms of discrimination.
https://asiatimes.com/2021/09/toward-an-inclusive-ai-future-for-women/
Read the full article by clicking on the above link
Problem:
AI has become an all-pervasive technology that penetrates businesses and societal landscapes. Humans have their own set of biases, and they design AI systems. When algorithms get applied to social and economic quandaries, it may lead to discrimination. In decision-making, it might cause algorithmic harm to marginalized communities or deprive them of opportunities.
A machine can process and analyze large volumes of data, but if the data is flawed with gender stereotypes, the results will reflect the bias.
Examples:
Amazon had to shelve its AI recruiting tool because it failed to rate candidates in an unbiased manner and began to play down women’s résumés. Women’s World Banking found that credit-scoring AI systems commissioned by global financial service providers resulted in the exclusion of women from loans and other financial services.
Short term Solutions:
- The main solution lies in collecting reliable, accurate, and continuous data. The questions for the data collection need to be flawlessly framed.
- It is the responsibility of social leaders and AI developers to advance gender equality.
- Lacuna Fund-a a multi-stakeholder collaborative of technical-experts, thought-leaders and end-users aims to make AI equitable by creating datasets that are representative of all classes.
Long-term solutions
- A heterogeneous workforce for identifying biases, solving issues, and asking the relevant questions .
- Including women in the decision-making process.
- Increasing the participation of women in AI-related fields
- Better representation of women in STEM(Science, Technology, Engineering, and Maths) by encouraging them to pursue these fields.
- Gender-sensitive governance structures and policies for responsible AI.
- Executive Management in companies should work towards developing equitable AI.
- The Algorithmic Justice Leagueand the firstgenderless voiceare two initiatives that are working for making AI less biased.