Greater gender diversity vital for AI to serve needs of society
Only 22% of AI professionals globally are female, according to the Global Gender Gap Report 2018, produced by the World Economic Forum (WEF).
Russia leads globally in gender equity in professional and technical fields, educational enrolment and health and survival. Lower scores in the realms of economy and politics drop its overall rank to 75th in the world. A Women in Science report from UNESCO found that 40% of researchers in Russia are women, against a global average of 28.8%.
Dorogush has no idea why her country performs strongly in terms of women in technology, except for its achievement of full gender equity in access to and participation in education.
“Higher education is super-important to providing many women with the skills to enter tech fields.” While there are no longer barriers to entering universities, it is crucial to make maths and science more popular among women, so that more enter the tech field, she says.
Once there, they find the tech environment more gender-friendly than most. “It is a good area for women. Maybe because it’s a sector where a lot of people are very well educated, and where there are fewer stereotypes.” What matters is a person’s skills set. “It is very black and white,” Dorogush explains, with few grey areas for prejudice to sneak in.
Still, the global picture is not encouraging.
A gloomy global picture
The WEF’s annual study benchmarks 149 countries on progress towards gender parity in four areas – economic participation and opportunity, educational attainment, health and survival, and political empowerment. The latest report, published in December, found that the global gender gap had slightly eased but fewer women were participating in the workforce. The world’s most gender-equal country is Iceland.
The latest edition also probes skills gaps related to artificial intelligence (AI), in partnership with LinkedIn. The analysis found that only 22% of AI professionals globally are female. The gap does not seem to be closing, with men and women adding AI skills at a similar rate.
“The Forum suggests that the rise of new technologies across a range of industries may, in fact, play a role in exacerbating persistent gender gaps,” writes Sue Duke, LinkedIn’s senior director of public policy, in a short background article for the WEF report.
Rapid technological changes are reshaping the global economy, and the field of AI is one of the main drivers of labour market transformations. As demand for AI skills expands across a growing number of industries, there is a risk that it could perpetuate – and even widen – the sort of gender and equity gaps that often impact on the technology sector.
“Some analysts contend that artificial intelligence itself could magnify inequality across a range of contexts, as algorithms often reflect the implicit biases of their creators,” says Duke. New AI systems, the report says, must serve the needs of society at large.
The research found that, as with the overall workforce, female AI professionals are more likely to work in “traditionally female” industries such as the non-profit, healthcare and education sectors. In general, women are likely to work in the use and application of AI, and men are more likely to work in development of the technology itself.
“In short, women are ‘growing but not gaining’ when it comes to AI skills,” writes Duke, and gender imbalances are likely to persist. “The far-reaching impact of artificial intelligence suggests that there are both equity and ethical imperatives to addressing the shortage of women in developing AI and other emerging technologies.”
Tackling the skills need
There is a clear role for higher education – and, looking at Yandex, also for the tech sector – in providing women with skills that can expand their participation in AI development. And there are lessons to be learned from Russia’s achievements in educational access.
Dorogush graduated from the faculty of computational mathematics and cybernetics at Lomonosov Moscow State University – a top faculty and university in Russia. “I was very lucky to have chosen a good place to study,” she says, that also encouraged women into the field.
After graduating, Dorogush studied for a three-year masters through the Yandex School of Data Analysis. It is, she says, “one of the top places in the world to study machine learning and definitely the top place in Russia”, recognised by the likes of Oxford, Harvard University and the Massachusetts Institute of Technology.
She then worked for various companies – ABBYY, Microsoft Bing and Google – before joining Yandex in 2015. Initially employed as a software engineer, she is now head of machine learning systems and is leading efforts to develop CatBoost, a state-of-the-art machine-learning algorithm that is available as an open source library.
Like some other tech companies, Yandex straddles the worlds of business and academia, especially in research, partnering with top institutions such as the Moscow Institute of Physics and Technology and the Higher School of Economics.
It has specialists in many disciplines, and over and above working on products, some also lecture and train. The Yandex School of Data Analysis has been running since 2007 and offers free courses for graduates or senior students who can later work for Yandex or other companies. It also has schools for project managers, user interface developers and others in tech and AI.
Yandex runs Russia’s largest technology conference as well as scientific conferences on machine learning, and seminars, lectures, workshops and master classes.
One of Europe’s biggest internet companies and Russia’s most popular search engine, Yandex develops intelligent products powered by machine learning. It has, for instance, produced a Russian-speaking intelligent assistant and a driverless car prototype. In 2009 Yandex developed and implemented its own machine learning method, MatrixNet, the predecessor of CatBoost.
Since being formed as a company in 2000 with 25 staff, three years after launching the web portal Yandex.ru, Yandex now employs nearly 10,000 people and has 30 offices worldwide. It is the largest technology company in Russia, the largest search engine in Russian, and the world’s fifth largest search engine after Google, Baidu, Bing and Yahoo!
Barriers to transformation
While Yandex has a high proportion of female employees for a tech company – more than a third of staff are female – only 19% of technology roles are filled by females.
The WEF-LinkedIn study found that Germany has one of the largest AI gender gaps – 16% of the AI talent pool is female – along with Brazil, Mexico and Argentina. The smallest AI gender gaps are in Italy, Singapore and South Africa, where 28% of the talent pool is female.
The gender imbalance has remained constant over four years, suggesting “a hardened talent gap that will require focused intervention”, says the Global Gender Gap Report 2018.
Aside from gender skews in ‘hard’ and ‘soft’ areas of AI, women in AI are less likely to be in senior roles and “less likely to gain expertise in a number of high-profile, emerging skills”.
Barriers to women continue to exist in tech as across most areas of life, Dorogush agrees, “as we continue to live in a patriarchal society”. But again, she believes tech and AI is exactly the place where there can be less of this – “it is a good area for women to be in” – so long as the number of women accessing the sector can increase, through education.
She argues that gender balance, reflected in equal numbers of women and men, should be the aim in all areas of work, including tech. But while balance is good, “it is important not to overdo the numbers” – quality should not be compromised during efforts to raise women’s representation, as this would undermine the long-term push towards gender equality.
As a leader in a major tech company – as one of the role models she believes are key to tackling the lack of women in tech – Dorogush understands that she can be a transformative leader.
Her way of trying to do this is, quite simply, to be scrupulously fair in dealing with people around her, to create an environment of honesty and trust. The tech sector is more impartial than most in terms of salaries and opportunities for men and women, and this level playing field should be extended to every aspect of the environment, for women to thrive.
The WEF reports suggest a more proactive approach in a context where gender gaps within the AI talent pool reflect broader gender gaps in science, technology, engineering and maths fields, across industries and in the acquisition of emerging skills.
“The data demonstrates a persistent structural gender gap among AI professionals, with well differentiated career trajectories taken by men and women in today’s labour market.
“Such figures should act as an early warning system to industries looking to achieve gender parity,” says the report. “Effective reskilling interventions and tangible job transition pathways will be key to narrowing these emerging gender gaps and can pave the way to reversing such trends.”