September 11, 2019

July 5, 2019

Please reload

Recent Posts

The Impact of Gender/Sex on AI

December 23, 2018

 

The mission of the iGIANT® is to accelerate the translation of research into gender/sex-specific design elements including for the IT sector which engages artificial technology (AI). As an iGIANT Scholar-in-Residence, I prepared a brief report examining the impact of gender/sex in this area which I have attached below:

 

-----------------------------------------------------------------------------------------------------------------------

 

According to journalist, Rachel Crowell, as of 2017, only 17% of women study computer science in college and beyond; however, women and men both use computers and AI technologies in their daily lives. While women consume media and technology, gender/sex-related discrepancies in technology limit female engagement in the field. The following issues need to be addressed:

 

1) The female voice is the default for the most popular figures like Siri, Cortana, and Alexa because most people favor computerized women’s voices over men’s. However, while many of these voices are gendered as female, AI assistants often have difficulty understanding women users’ voices. AI assistants are trained to recognize and respond to their users by being exposed to thousands of samples of voices. Most of these voices are from white men; therefore, assistants like Siri or Alexa may not recognize certain words spoken in higher-pitched, more feminine voices. 

 

2) Software tends to associate certain images such as food, makeup, childcare, clothing, etc. to women. This bias can limit the number of career results, travel results, or other non-domestic social media advertisements and suggestions for women. Women and men are shown different advertisements and content as they scroll through their social media feed even though businesses could benefit from a larger, more gender-diverse pool of viewers to purchase their products. 

 

3) AI algorithms are designed to work out problems, replicating human patterns of thinking and behavior. The problem is that men and women often solve problems differently, so computer science is often skewed toward a stereotypically masculine mindset:

  → Women tend to solve problems by collecting information first then they act upon their findings. 

  → Men tend to act before assessing information, learning and readjusting along the way. 

 

4) Translation apps do not offer the user the opportunity to specify a gender. When translating even simple sentences, Google Translate assumes that the singular user is male; therefore, women may encounter frustration or confusion in these encounters. 

 

5) According to scholars Charles Hill, Maren Haag, Alannah Oleson, Chris Mendez, Nicola Marsden, Anita Sarma, and Margaret Burnett, women and men interact with computers in different ways including the following: 

  → Women are more risk-averse when learning new technologies. 

  → While women generally “tinker” with their technologies less than men do, women are more                    reflective and tend to learn more than men when they do an experiment in this way. 

  → Technology users, both male and female, benefit when computer scientists create AI that has                   male/female personas and can account for multiple problem-solving techniques. 

 

6) Women generally use technology for a practical purpose or to assist with tasks, whereas men are more often motivated by entertainment. 

     → Because of this practical approach, women are less tolerant of difficult-to-use interfaces or                      confusing programs and prefer cleaner, easily navigable programs.

 

The purpose of this overview is not to indulge essentialist comparisons of women and men, especially since such stereotypes often lead teachers to guide young women away from STEM fields in high school and beyond. However, we do need to address the existing gaps in the field that continue to dissuade women innovators and clients. 

 

In conclusion, the next time you use a piece of artificial technology, whether that be your smartphone or a gaming system, ask yourself how inclusive that technology truly is. Consider the size of the phone in your hand or in your pocket; is the device designed with women's bodies and lives in mind? With these questions, you will be engaging in a larger discourse of equitable access, and you can help change the face of AI to be more inclusive of all users. 

Share on Facebook
Share on Twitter
Please reload

Please reload

Archive
  • White Facebook Icon
  • White LinkedIn Icon
  • White Twitter Icon

© 2017 by PRITI KAUR.