Ying Liu is the head of AI R&D at the London-based startup TG0, and a pioneering researcher in human-computer interaction. Here she argues that progress will be hindered if the sector does not embrace diversity, particularly at the top.
The pace of AI-fueled innovation has accelerated rapidly in recent years, matched only by the demand for experts to oversee such growth. Consulting firm Bain & Company estimate that AI-related jobs have surged by 21% each year since 2019, with the number of qualified candidates failing to keep pace. The UK could see talent shortfalls of more than 50% by 2027, with just 105,000 AI workers available to fill the 255,000 AI jobs.
Faced with such numbers, I find myself wondering why the number of women employed in AI remains so low. While there are some trailblazers, such as Mira Murati, the former CTO of OpenAI, women only account for just 22% of AI talent globally and less than 14% of senior executive roles in AI.
As head of AI R&D at TG0, I have been taken aback by the gender imbalance in AI spaces. This isn’t just about equality. If we don’t give women the same access to the networks and discussions where AI technology is developed, the world will not benefit from its full potential. This gender gap represents a significant loss of potential talent.
Bias built in
It’s well established that many young women are put off studying STEM from an early age. PwC has found
Only 30% of girls study STEM subjects at university and only 3% say a tech career is their first choice. Over a quarter say they’ve been put off a career in technology because it’s too male dominated.
There’s work to do to show younger women and girls that this is a viable career choice for them too. I myself came to coding late, although I loved physics and maths at school. Now, I try to be a visible leader with career talks and mentorships through the Women Coding Community, which I co-founded.
Because diversity matters in tech, particularly when it comes to AI. Diverse teams are more likely to challenge the status quo, think outside the box, and advocate more inclusive practice. Without diversity, we won’t build AI systems that work for all our society.
This has been recognised by giants such as IBM. In a recent poll run by the company 73% of business leaders believe that increased female leadership in the sector is important for mitigating gender bias in AI and 74% say increased female leadership is important for ensuring the economic benefits of AI are equally felt across society.
Indeed, we’ve already seen examples of this – systems designed for facial and voice recognition technology that misidentifies women and individuals from ethnic minorities. Job recruitment algorithms and a financial lending algorithm that offered lower credit limits to women than men.
It’s about influence
AI innovation is happening at speed but we need to slow down enough to include women in the revolution. AI could become the most pivotal tool for advancing equality or set us back decades. I’m incredibly proud in my role as head of AI R&D to meet some incredible women working in this field, but we need more – not only working with AI but leading the work with AI.
Diversity isn’t just about the numbers in a room but also influence.
If we can encourage young women to go into technology, we need to ensure they stay, by nurturing their talent and supporting their growth into leadership.
In technology as a whole, it’s currently estimated that 50% of women leave the sector before they are 35
Of those that do progress, only 5% will ever hold leadership positions.
Diverse leadership teams have been found to deliver up to 30% better outcomes in innovation and ethical decision making, something that will be sorely needed as we expand the use of AI.
Women bring unique perspectives that balance innovation and the fail fast, fail often mentality with compassion and efficiency.
AI has the potential to solve some of the world’s greatest challenges, but women need to be equal partners. That won’t automatically solve all of the challenges facing AI developers of course, but it will go some way towards ensuring the future is one of innovation, fairness and mutual benefit for all.
