Busting Bias in Artificial Intelligence and Reclaiming Jobs Campaign
STOP: Slow down your thoughts
THINK: Are you conforming to a stereotype?
RECONSIDER: Your decision
"Do not judge me by my skin colour, my hair colour, my body size and my outer beauty. If you do, you will miss entirely who I am."
The Problem:
A new study published in the journal 'Science' shows the stereotypes we unconsciously hold about women and minority groups are replicated in the fastest growing area of technology, Artificial Intelligence (AI). Moreover, a study conducted by the Girl Guides Association showed that girls and young women said they held themselves back due gender stereotypes, sexism, and anxiety about how they looked. There was a significant increase this year of mental illness among young girls, also due to increased pressure on how they look. In addition, Mashable reported on a study which showed girls were affected by stereotypes by age six.
What we want:
Decades of research such as this one show that stereotypes and unconscious biases reduce self-esteem, motivation, and intellectual performance. With the use of AI (computer systems that think and act like humans) on a rapid increase, we are urging the Government to launch a jargon-free publicity campaign to help the layperson spot biases and override them.
Key points:
+ Unconscious bias is when we make snap judgements and assessments about people, especially those who don't look like we do. It happens to us all, however, especially when we are tired or rushed.
+ We are becoming increasingly reliant on AI to make important decisions for us, such as in recruitment selection.
+ The biases of the individual coders get passed on to AI.
+ While humans can counteract their learned behaviours, the computers can't.
+ This has resulted in the reinforcement of sexism, racism and other forms of discriminatory outcomes.
The Facts
- There were instances where algorithms were only showing highly paid roles and jobs in STEM to men.
- Robot judges were used to shift through 6000 applicants in a beauty contest assessing for facial symmetry and wrinkles. Due to the data, it was provided it didn't flag up anyone with dark skin as a winner.