Is AI a panacea, or an existential threat? The future of artificial intelligence divides opinion.
It’s becoming increasingly clear that the reality is more nuanced than these binary positions and the pace of change is humbling. What is clear is that the debate becomes more complex when you look at it through the lens of gender equality.
Without careful regulation, deliberate management, and inclusive design, with the scale and rate of progress we’re seeing today, AI will widen the gender gap, not close it.
Because AI is not inherently neutral. It reflects the biases of the world and right now, it is amplifying gender inequality.
From hiring algorithms that filter out women, to AI models which undervalue and misinterpret women’s medical conditions, or facial recognition systems that misidentify them, gender bias in AI is not theoretical—it’s playing out in real time.
Amazon famously abandoned an experimental AI hiring tool after it was found to penalize résumés that included the word “women’s” or were from all-women colleges. The algorithm had “learned”, from biased historical data, that male candidates were preferable. A recent audit of open‑source LLMs found they tend to prefer male candidates—especially for higher-paying roles, directly mirroring occupational stereotypes.
This isn’t just happening in recruitment. Gender gaps in healthcare, long fueled by male-centric data, are now being replicated by AI systems trained on the same flawed inputs—leading to diagnostics and treatments that disproportionately fail women. A large U.S. study published in Radiology found that an FDA‑approved AI mammography algorithm was 50% more likely to generate false‑positive results for Black women compared to white women.
And in fintech, credit-scoring algorithms are penalizing women. A 2019 investigation by the New York Dept of Financial Services into Apple Card’s credit algorithm found that women were given significantly lower credit limits than men — even when they had better credit scores.
A 2021 working paper from the National Bureau of Economic Research (NBER) analyzed data from fintech lenders using algorithmic credit models advertised as “objective” and “bias-free,” still charged higher interest rates to women compared to men with identical credit profiles.
Finally, when it comes to large language models like ChatGPT, the challenge is equally complex. These tools reflect the content they’re trained on—vast swathes of the internet shaped by historical and societal bias. A 2023 Stanford study found that language models frequently default to stereotypes: associating nurses with women, leaders with men, and codifying assumptions that subtly reinforce inequality.
The biases found within LLM outputs are perpetuating the problem, by deterring women from using the tech in the first place. 50% of men use generative AI tools, compared to 37% of women. This leads to, as a Harvard Business School paper suggests, “systems trained on data that inadequately sample women’s preferences and needs”.
If we don’t correct for these embedded biases, we risk baking discrimination into the foundation of our digital future.
To truly harness the potential of AI for good, we must ensure it is shaped by a diversity of perspectives, grounded in empathy and equity. In other words: representing and serving the needs of all humans, not just those in positions of power.
So how do we change the story?
Let’s start with the numbers: women and gender minorities make up just 22% of AI professionals. In OECD countries, fewer than one in three AI-related graduates are women. Among AI-specific PhDs, that number drops to 18%.
In the startup ecosystem, the picture is even starker. In the UK, according to the Alan Turing Institute, female-led AI startups receive just 0.03% of total AI equity funding— with the average deal size coming in at six times less for women (£1.3m) than for male teams (£8.6m). The number of AI startups with all women teams is small, and when it comes to innovation ownership, only 2.5% of AI-related patents globally list a woman as the lead inventor (PwC/WEF, 2023).
Would more women pursue AI if they had equitable access to capital? If they saw themselves represented in leadership and technical roles? If they weren’t building systems within structures rigged against them? I’d like to think so.
Workplace culture remains another barrier. According to McKinsey (2024), 40% of women in AI roles report feeling excluded from key projects or decision-making—and nearly one in three are considering leaving the field altogether.
That’s why investing with a gender lens matters.
Gender-lens investing channels capital to women-led or gender-diverse AI teams. It expands access to funding, creates inclusive leadership pipelines, and ensures more voices are shaping the technologies that will shape our future.
This strategic approach addresses systemic inequities, promotes innovation, and strengthens both business and societal outcomes. It ensures that AI technologies are developed and deployed in ways that are more inclusive, less biased, and better aligned with the needs of a diverse global population, ultimately leading to more sustainable and equitable futures.
Diverse teams are also better for business: McKinsey data shows that companies in the top quartile for gender diversity are 25% more likely to financially outperform their peers.
Beyond moving capital to more gender diverse companies and teams, investors can use gender-focused frameworks and due diligence tools to assess how AI systems affect different genders, ensuring that potential harms are identified and mitigated before scaling up.
We cannot build ethical AI without diversity. We cannot build inclusive AI without women and diverse leaders at the table. And we cannot change the future without changing where the money flows.
If we want AI that works for everyone, we must invest responsibly—and that starts by investing in everyone.
