Are AI assistants teaching girls to be servants?
FYI, this story is more than a year old
Have you ever interacted with a virtual assistant that has a female-based voice or look, and wondered whether there are implicitly harmful gender biases built into its code?
It turns out the United Nations’ EQUALS Skills Coalition has done some research and arrived at that very conclusion – there’s a gender divide in our virtual assistants and it's not good for our women. Think Amazon’s Alexa, Apple’s Siri, and even the default voices when we navigate the streets through Google Maps.
The study, titled I'd blush if I could: Closing gender divides in digital skills through education is named after one of Siri’s previous responses when users insulted ‘her’. Now, ‘she’ just says, “I don’t know how to respond to that”.
The study assets that Siri’s ‘female’ subservience and servility shows how gender biases are coded into technology products and the tech sector – and how they are affecting digital skills education.
“Today, women and girls are 25% less likely than men to know how to leverage digital technology for basic purposes, four times less likely to know how to programme computers and 13 times less likely to file for a technology patent. At a moment when every sector is becoming a technology sector, these gaps should make policy makers, educators and everyday citizens ‘blush’ in alarm,” the study cautions.
It is a somewhat brazen use of female subservience as an alarmist tool to get people thinking about the digital skills gender gap, but it does make a serious and valid point – and those in industry can see the reality.
"The people behind the majority of today's technological advances ' the workers creating the algorithms ' are predominantly white and male. And even more importantly, when white male coders assemble data for chatbots, machines are likely to perpetuate inequities found in the real world,” comments Skillsoft CMO Tara O’Sullivan.
“They are prone to hard code their own subconscious bias about race, gender and class into algorithms that are designed to mirror human decision making. This has the propensity to amplify existing stereotypes and create a stronger association for male and female-oriented images, behaviours and careers.”
She says that coders need to be trained to recognise their own biases – and that there needs to be a bigger effort to get women and girls in STEM.
The report adds, “UNESCO estimates that men are around four times more likely than women to have advanced ICT skills such as the ability to programme computers. At the frontiers of technology, the gap becomes an ocean: Across G20 countries just 7% of ICT patents are generated by women,12 and the global average is even lower, at 2%. Recruiters for technology companies in Silicon Valley estimate that the applicant pool for technical jobs in artificial intelligence (AI) and data science is often less than 1% female.”
What’s more, the number of female computer information science majors has dropped both abroad and right here in New Zealand.
The digital skills gap begins in primary school, but it is well pronounced in secondary schools. “Girls tend to opt out of STEM subjects earlier in secondary school than boys,56 meaning that they are increasingly less likely to pursue technology-related studies as they move through secondary school and into higher education,” the study says.
It is important to close that skills gap – anything that helps women and girls to develop digital skills has benefits for stronger women, stronger families, stronger communities, stronger economies, and better technology, the study posits.
The study points to how women in Brazil and India have improved their communication with local government by using technology. The Women-gov project has helped to establish women-run, internet-connected community information centres for social assistance.
So how does the world go about closing the digital skills gender gap? The study offers 18 recommendations to help women and girls develop their digital skills. Those recommendations include: Sustained, varied, and life-wide approaches to digital skills that start early.
“Motivating girls to become skilled ICT users is likely to require new and varied learning pathways. The ‘one-size-fits-all’ approaches attempted in countries have tended to magnify rather than reconcile existing gender inequalities. Recently, however, models of learning techniques to better accommodate girls and women have been coming into focus.”
Incentives, targets and quotas could also help to boost the number of women who specialise in ICT. Scholarships should be formed with caution: “For example, funding opportunities labelled as ‘diversity scholarships’ to technology programmes may be less effective at boosting women’s self-confidence and sense of accomplishment than academic scholarships. Such wording may reinforce women’s outsider status and make them question whether they have the aptitude to succeed in a field currently dominated by men.”
Other recommendations include the recruitment and training of gender-sensitive teachers; promoting role models and mentors; involving parents in digital skills education; and the support of technology autonomy and women’s rights.
“Solutions to turn the tide will require a critical mass of efforts; the problem is too deep and multifaceted to address with singular actions – no matter how ambitious,” the study concludes.
“This paper outlined some of the most promising approaches, scanning the world for policies and programmes that are ‘un-gendering’ technology and helping women and girls to develop their digital skills and gain confidence in gender-responsive learning environments. Education has a key role to play in this process. It is where expectations are forged and competencies are cultivated.”