AI represents a quantum leap in technology – but are women being left behind?

Praised for its potential to elevate business productivity and solve a range of societal issues – from diagnosing diseases to enhancing education – artificial intelligence (AI) is set to greatly impact our future.

As its adoption accelerates, however, there are signs that AI may not be for the benefit of all. While AI-led tools can liberate us from certain tasks, the tech’s current trajectory could have equally oppressive results. And when it comes to the advancement of women in tech, there’s a real concern that AI could represent a step backwards.
 

A male-dominated AI investment space

Breakthroughs in AI have led to significant investment in the tech, which Goldman Sachs predicts could reach $200 billion globally by 2025. Amid this financial boom, though, women’s representation in AI entrepreneurship is minimal. Female-founded companies accounted for a mere 2% of AI startup deals over the last decade, according to a recent report by the Alan Turing Institute. And for all-female teams, this figure is less than half a per cent (0.4%).

Issues concerning AI and gender equality go beyond venture capital, and are reflective of broader underrepresentation in emerging tech fields. Earlier research shows that women make up barely a fifth (22%) of AI and data science professionals in the UK, and constitute only 8% of researchers contributing to leading machine learning conferences. The result is a cyclical pattern of exclusion: a lack of female representation in AI only compounds barriers to access for women who want to pursue careers or education in wider tech fields, which could increasingly require AI competencies.
 

A lack of diversity could perpetuate bias

What happens when technology is built from the ground up by an unrepresentative body of people? In all likelihood, the final product will miss out on crucial input and questioning that makes it representative of the society it’s designed to serve. And those missing from AI research and design are the most likely to suffer as a result.

When AI teams are homogenous, there’s a greater risk that human biases will creep into AI systems and be amplified at scale, potentially discriminating against women and other marginalised groups. A study conducted by computer scientist Joey Buolamwini found that facial recognition systems performed substantially better on male faces than female faces. Moreover, while lighter-skinned men experienced error rates of no more than 1%, this leapt to 35% for darker-skinned women.

The exclusion of women in AI will ultimately result in a loss of innovation and creativity in the tech’s development, leading to missed perspectives and experiences. Gender-diverse teams are needed to mitigate the risk of embedding existing biases into tools that could potentially influence the thoughts and opinions of millions.
 

Challenging gender imbalance in AI

Democratising AI will not be an overnight process – and existing social biases may need to be addressed first – but there are some steps organisations can take together to close the gender gap in AI and create a more equitable future:

• Collaborate to innovate: Organisations and key decision-makers should take a multi-stakeholder approach in a bid to draw more women into AI fields. This includes building greater relationships with diverse tech groups, networks and other entrepreneurial communities to facilitate broader representation and investment opportunities in AI.
• Diversify hiring processes: Employers should endeavour to eliminate unconscious biases that may exist in their recruitment processes. Job descriptions must use more inclusive, gender-neutral language that encourages women to explore opportunities in AI. Digital recruitment tools and algorithms – especially those powered by AI and machine learning – should also be as transparent as possible.
• Create an inclusive culture: Open and equitable hiring must be complemented by workplace policies that engender a sense of belonging and aid career progression. Consider how women can be better supported in their tech careers. For example, through flexible working and parental leave policies, or fair and transparent pay.

AI’s gender imbalance has negative outcomes for us all, and it’s imperative that women are given the platform they deserve to author the future of AI – one that benefits both business and society.

Get your copy of our latest DE&I report today for more diverse hiring and workplace inclusion recommendations.