Q&A  | 

Saniye Gülser Corat, UNESCO Gender Equality Director

"Amazon’s AI recruiting software was found to downgrade résumés that contained the word ‘women’s’".

Tags: 'digital gender gap' 'gender equality' 'ICT gender equality' 'Saniye Gülser Corat' 'UNESCO'

SHARE

Saniye Gülser Corat is the Director of the Division for Gender Equality in the Office of the Director-General at UNESCO Headquarters in Paris, France. Before joining UNESCO in September 2004, she pursued a diversified career as an academic at Carleton University in Canada and a senior international development advisor to several international organizations, planning, managing, monitoring, evaluating projects and programmes in areas ranging from education, social and economic development, institutional strengthening, gender equality, technology transfer to local governance and climate change in over 30 countries in Asia and sub-Saharan Africa. She also has private sector experience as the Chief Executive Officer (CEO) of an international development consulting company in Canada for almost 10 years.

Can you give us an overview of your work?

I am the Director of the Division for Gender Equality at UNESCO, since September 2004. In 2007, UNESCO’s Member States decided to designate gender equality as one of the two global priorities of the Organization. Since then, I strove to make UNESCO not only an intellectual leader but also a role model through its initiatives and actions to advance gender equality. 

To achieve this goal, we generate knowledge on gender equality issues through research and flagship publications and publish studies on topical issues such as the gender divide in digital skills, gender bias in artificial intelligence (AI) and STEAM education for girls. Building on such research, we develop and implement programmes that respond to the needs of women and girls around the world. To address specific inequalities between women and men, UNESCO also strikes effective and innovative partnerships.  

UNESCO was the first UN agency to develop a customized Priority Gender Equality training programme and to make this training mandatory for all staff as early as November 2005.

UNESCO is also recognized as a model UN agency for creating a gender responsive workplace and for achieving gender parity in its staffing, including at decision-making levels, improving the representation of women at Director and above levels from 9 % in 2004 to 50 % in 2017.

What do we understand by the term: digital gender gap?

The digital gender gap or divide refers to the differences between men and women and between girls and boys in their ability to access and use digital technologies and participate fully in the online world.

Worldwide, women and girls are disproportionately affected by the lack of access to information and communication technologies (ICTs) and lack of digital skills to use them effectively.

What are its causes?

According to research conducted in 2019 (EQUALS Research Group), the digital gender divide generally occurs because of one or more of the following factors:  

  1. Broadband internet infrastructure is not ubiquitous, especially in rural areas.
  2. The cost of IT hardware and software is often prohibitive for women since they generally have less purchasing power and financial independence than men, especially in less developed countries
  3. Women tend to have lower levels of digital literacy than men and often lack even basic digital skills to use the technologies effectively
  4. Women tend to be less interested in ICTs and tend to underestimate their benefits and utility. This might be because of their lack of exposure to these technologies, and because relevant content, in accessible languages and formats, is unavailable
  5. Online threats and cyber violence against women contribute to keeping women away from ICTs
  6. Social norms, stereotypes and cultural constraints may make it difficult for women to access digital technologies or receive the required training to use them effectively.  

How big is it and how is it evolving?

The digital space is becoming more male-dominated, not less so.

Looking at the number of men and women using the Internet worldwide, the International Telecommunications Union (ITU) concluded in 2019 that the digital gender gap is growing. While women were 11 per cent less likely than men to use the Internet in 2013, this has increased to 17 per cent in 2019. Over half the total global female population (52 per cent) is still not using the Internet today.

The digital skills gaps also appear to be growing. Indeed, women are globally less likely to know how to operate a smartphone, navigate the internet, use social media and understand how to safeguard information in digital mediums – abilities that underlie innumerable life and work tasks and are relevant to people of all ages. According to cross-national skills assessments, women in numerous countries are 25 per cent less likely than men to know how to leverage ICTs for basic purposes, such as using simple arithmetic formulas in a spreadsheet. Further along the skills spectrum, the divides grow wider. 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 per cent of ICT patents are generated by women, and the global average is even lower, at 2 per cent.

Can you talk to us about the ICT gender equality paradox, by which advanced digital skills education is an ‘outlier’ with regards to its association with gender equality?

It refers to the surprising lack of a direct relationship between gender equality levels and the proportion of female students pursuing advanced-level digital skills.

In other words, countries with higher gender equality levels, in Europe for example, do not always have a higher proportion of women pursuing ICT-related diplomas.

On the other hand, countries with lower levels of gender equality, such as Arab region countries, report relatively more women completing ICT programmes than do more gender-equal countries. That is why advanced digital skills education is considered an ‘outlier’ with regard to its association with gender equality.

It also suggest that while gender specialists do not often look to the Arab region for policy lessons in how to achieve gender equality, the narrow area of advanced ICT education may constitute a notable exception.

As a caveat to these findings, it is worth noting that many female students pursuing advanced-level digital skills in Arab countries have difficulties putting their skills to economic use in the workforce.

Which are the social and economic consequences of the digital gender gap?

The proliferation of digital technology and digital services has made digital skills a prerequisite for full participation in society.

As such, the digital gender gap puts women at risk of being left behind in all realms of economic, political and social life.

Considering women make up over half the world population, this is not only worrying for women themselves, but has dire consequences for economic growth and societal well-being as a whole.

One specific example of the consequences of the digital gender divide is the lack of diversity in the technology sector. Recruiters for technology companies in Silicon Valley estimate that the applicant pool for technical jobs in AI and data science is often less than 1 per cent female.

Worldwide, only 22 percent of AI professionals are female while women make up only 12 percent of AI researchers.

This has serious social and economic consequences. First, because as stated by the OECD in 2019, “inventions arising out of mixed teams, or women-only groups, appear to have wider technological breadth (and may therefore be more economically valuable) and higher impact from a technological viewpoint than those in which only men are involved”. Secondly, because as shown in UNESCO’s I’d Blush If I Could report, technologies generated by male-dominated teams and companies often reflect troubling gender biases.

Amazon’s AI recruiting software, for example, was found to downgrade résumés that contained the word ‘women’s’, as in ‘women’s chess club captain’, because it had been trained on men’s résumés.

Can you explain to us a case of gendered AI technologies and gendered biases coded into tech?

An example of gendered AI technologies are digital voice assistants. Today and with rare exception, most leading voice assistants are intentionally humanized and more often than not exclusively female or female by default, both in name and in sound of voice.

Amazon has Alexa (named for the ancient library in Alexandria), Microsoft has Cortana (named for a synthetic intelligence in the video game Halo that projects itself as a sensuous unclothed woman), and Apple has Siri (coined by the Norwegian co-creator of the iPhone 4S and meaning ‘beautiful woman who leads you to victory’ in Norse).

In addition to anthropomorphizing digital voice assistants as female, overwhelmingly male engineering teams in technology companies may also code gender biases in these devices. Companies like Apple and Amazon, for example, originally designed their feminized digital assistants to greet verbal abuse with catch-me-if-you-can flirtation.

In 2017, Quartz investigated how four industry-leading voice assistants responded to overt verbal harassment and discovered that the assistants, on average, either playfully evaded abuse or responded positively.

It projects a digitally encrypted ‘boys will be boys’ attitude.

At UNESCO, we believe that the structural sexism of feminized digital voice assistants cannot be solved by simple software updates but requires instead a thorough rethinking of the enterprise of gendering machines.

Why should we pay attention to whether AI technologies are gendered or not?

It is important to remain vigilant vis-à-vis the gendering of AI technologies as it has the potential of spreading and reinforcing harmful gender stereotypes in our societies. For example, the 2017 Quartz study of how four industry-leading voice assistants responded to overt verbal harassment concluded that the evasive and playful responses of feminized digital voice assistants ‘reinforce stereotypes of unassertive, subservient women in service positions… [and] intensify rape culture by presenting indirect ambiguity as a valid response to harassment.’ The four voice assistants studied failed to encourage or model, let alone insist on, healthy communication about sex or sexual consent. Their passivity, especially in the face of explicit abuse, reinforces sexist tropes.

The gendering of AI may therefore further stigmatise and marginalise women from societies and may even offset the considerable progress societies have made towards gender equality in offline environments.