Skip to content

Is AI Increasing The Bias Against Women In The Workplace?

Less women are choosing to use AI in professional life, research shows, but for those that do, there exists a gender bias that holds them back in both physical and virtual spaces. Image by Alemedia.id via Canva
  • Women are more sceptical and less inclined to use artificial intelligence due to concerns around the ethics of using it
  • With the lack of female representation in STEM roles, businesses are worried women will be left out as AI use in the workplace continues to increase
  • AI is trained from data online, and with so many misogynistic content, the AI picks this up and is therefore more likely to discriminate against women

The gender divide exists in almost every aspect of society. Turn to the workplace and you’ll see women receiving 12.8% less pay than men and typically all whilst doing the same job, far more men occupying C-suite roles (and more of them named “John” than women named anything!), and more men gaining promotion and progression all thanks to the other men hiring them.

At home the story is no different. Women and mothers in the US, for example, have been found to handle as many as seven out of 10 household tasks, with men taking on less than half of them, and studies reveal that women commonly work longer, and make more professional sacrifices in order to keep family life afloat than male partners do. 

The divide is even apparent in our leisure time. Look at hobbies – often something people use as escapism – and you find a 24% team sport gap between girls and boys, which reaches into professional sports too, as male sports players routinely get paid significantly more than women. 

Most gendered ideas of male and female roles in life, work (and sport) are typically entrenched in tradition – old ideas and values passed down and through society. Newer generations may have a more liberal take and equal stance, but ingrained habits are hard to shake. 

However, this gender inequality has now seeped through into never avenues too. Artificial Intelligence (AI) is discriminating against women, giving humans even more reason to question AI’s integrity, trust, and ethics. 

Maintaining ethics and integrity 

It appears that men seem to be slightly more enthusiastic about the increase in the technology than women, as Associate Professor Rembrand Koning at Harvard Business School found that women are adopting AI tools at a 25% lower rate on average compared to men. 

In the study, the researchers further found that women use ChatGPT less than men, and the share of women adopting AI tools was 10-40% less than men. 

The reason for this, they state, is due to the ethics of AI, as women are more concerned about where to draw the line. Old gendered ideas of competency between men and women are also a contributing factor, with a greater proportion of women fearing they will be judged by their work colleagues for relying on AI tools. 

Professor Koning notes that “Women face greater penalties in being judged as not having expertise in different fields. They might be worried that someone would think even though they got the answer right, they ‘cheated’ by using ChatGPT.”  

Likely, the male workers in this scenario held no such concerns.  

Unfair? Certainly. But, are these women onto something by not using AI? With so many relying on the tech to do everything from drafting emails or providing assistance in learning, to presentations, or even developing projects, they may be missing out on simple daily opportunities to develop skills and relationships. Research has found that using AI chatbots can make you stupider. 

As a result, businesses face a complex reality, where they may begin to miss out on major productivity gains if ChatGPT use remains high, as staff may fall behind in building valuable skills they may need in business.

And, for the women who continue to dismiss and resist generative AI, whilst they may keep a sharper skillset and a more ethical approach to work, if businesses continue to fall into gendered traps of hiring and promotion, they may be missing out on getting the best staff. 

Worth the risk? Perhaps not 

Alongside women being more hesitant about the ethical implications of AI, research finds they are also much more sceptical in general of using it than men. 

The study, authored by Assistant Professor Sophie Borwein and colleagues interviewed participants about their views on AI’s risks and benefits. The women were found to be more likely to express uncertainty and scepticism. 

When asked what the benefits of AI are, 6-7% of women said that they “don’t know”, and 2-3% of women said that “there are no benefits”. 

Here, gendered habits also come into play. Co-researcher, Assistant Professor Beatrice Magistro argues that this divide in opinion regarding AI is due to “exclusion and socialization”. Women have historically been excluded from STEM fields, which are more male-dominated, and men continue to dominate positions of power in technology firms. 

“Women’s under-representation in science, technology, engineering and mathematics fields means they may have less access to high-paying AI-related jobs and leadership positions,” she notes. “As a result, advances in AI could widen existing gender pay gaps.” 

STEM is leaving women out 

Investigations conducted by faculty at the London School of Economics and Political Science similarly find that with women not being represented in STEM areas, they could get left behind in an AI-driven economy, further compounding the gender divide.  

Already, women have been found to make friends at work differently compared to men, which can impact their progression. Men, research finds, take a more calculated approach to their work friendships, taking consideration into the department someone is in and its professional benefits before making a connection. Women’s networks are warmer, far wider but perhaps not as progressive career-wise. 

So, while women are already at a risk of being left out of beneficial work friendships, their roles might also become automated ,ore swiftly. A staggering 79% of employed women in the US work in jobs that have a large risk of automation, compared to the 58% of men. Already, many female-dominated industries are being replaced by AI systems. 

And for the minority of women working in STEM and AI-focused roles, the lack of female representation in these areas means they too could get left behind in the shift to an AI-driven economy.  

And the numbers of women entering such fields may not change any time soon. Research by Professor Paola Profeta from Bocconi University additionally found that women face more math anxiety in STEM entry tests than men do, making job entry all the more difficult. 

Would removing time pressures in exams help the gender divide to close? Perhaps. But here too gendered expectations can create barriers. King’s College London explain that teachers in mathematics expect more from female students than male students, placing more additional pressure on them to perform well. 

The AI discrimination against women’s aging 

Stanford University have found that there is an AI bias against older working women too.  

For the study, the researchers asked ChatGPT to produce resumes for hypothetical women. The AI portrayed them as younger and less experienced, however, when generating resumes for men, it gave older men higher ratings, even though they were based on the same initial information as the women’s hypothetical resumes. 

The reason for the difference? AI ‘learns’ by gathering information from online datasets, so essentially, whatever is online, it will learn from. If bias exists in histories, AI will adopt it. 

In an interview with UN women, Beyza Do?uç, an artist from Turkey, explains she encountered gender bias in generative AI when using it for researching a prompt for a novel.  

The AI made the doctor male and the nurse female, and after giving it more prompts, the AI always chose stereotypical gender roles for characters. When Do?uç asked why, the AI explained that it was because of the data it had been trained on. 

Do?uç says that “Artificial intelligence mirrors the biases that are present in our society and that manifest in AI training data.” 

And whilst bias in general exists across all areas of society, the internet, particularly in pockets of social media, can be a breeding ground of hate and misogyny. For example, an investigation conducted by Amnesty International UK found that 73% of Gen Z social media users have witnessed misogynistic content online. 43% of women online have been body shamed, 40% of women are subjected to sexually suggestive comments, and 44% receive unsolicited explicit images. 

As AI learns from the data it is presented with without judgement, this only fuels the discrimination and gender divide. For AI developers, this presents a significant problem.  

The algorithm is playing favourites – or perhaps the ones who designed them are 

The disclination against women isn’t only found in generative AI and AI models. It’s also found on popular social media platforms – not just in the hateful comments towards women, but within the actual algorithm itself. 

In 2025, following an update to LinkedIn’s algorithm, active users Jane Evans – a Creative Director and author who had built a following of 19,000 and Cindy Gallop – a Founder CEO with a near 150,000 strong following, noticed a significant drop in the level of engagement on their posts.  

The steep decline in reach prompted them to investigate, conducting an experiment to explore what would happen if men and women posted the same content, at the same time. Two male participants, who had a joint following of 9,400, saw their posts achieve significantly more reach compared to the two female participants. 

Gallop shared her findings with LinkedIn, who initially launched an investigation, but nothing came of it. 

They were not the only women to have noticed the shift in visibility. The experiment of women changing their information and even their profile images on LinkedIn to become men, and then watching how their reach and engagement improved became something of a trend amongst senior professional women, making the case for the existence if the bias harder to deny. 

In November of the same year, Kamales Lardi shared the results of her own experiment with her over 20,000 followers. 

“I changed my gender on LinkedIn to ‘male’ and my impressions increased by 421% in a few days,” she began in her post. Having adapted her image, changing her “profile demographic to male” and using the help of ChatGPT to help her mirror a more male tone, she saw an undeniable spike in her reach after experiencing the same sharp decline that Evans and Gallop had shared. Her post provided the evidence – screenshots of her posts’ impressions from before and after the switch. 

“We have known for a while that social platforms and AI models carry bias. But somehow, LinkedIn has escaped this wave of backlash, perhaps because many of us are dependent on the platform for our business and professional progression. But the numbers don’t lie and this is NOT OK,” she shared, calling for change. 

Her views carry weight. Backed by multiple business and tech qualifications from Durham University Business School, The London School of Economics, Harvard and Wharton, Kamales is CEO of Lardi & Partner consulting and an AI and digital transformation expert, having quite literally written the book on the topic, as well as presenting on it for TEDx.  

Going forward 

The response from LinkedIn was disappointing. A statement made in the aftermath of the outcry disagreed with the findings and insisted there was no bias present.  

All the founders of LinkedIn are male, the same people who control and manage the algorithm and ultimately can help to influence which profiles perform the best. Perhaps they could do with taking a leaf out of Amazon’s book? 

In 2018, Amazon decided to remove its AI recruitment tool, as it was proven to have a bias against women. Additionally, the system also penalized resumes which included the word ‘women’s’ in it, for example; a resume stating the applicant had been a ‘women’s chess club captain’. A positive step, but as LinkedIn is a far more active recruitment space, more needs to be done. 

When looking at ways to solve this AI discrimination issue, it isn’t as easy as simply changing how an AI learns or having more women in STEM positions. Misogyny is found everywhere: school, work, online, and even subtle mentions in literature, TV, and video games. 

AI takes data from all of these to learn, and while it sadly can’t be changed overnight, collaboration needs to happen to make the change. 

Like Professor Profeta notes, changing the exam formats to ensure women are given the same advantages as men will allow more females opportunities to enter STEM careers, such as AI development, or creating the platforms and services that go on to use AI tools and algorithms in their operations.  

With more women occupying the roles which have a say in artificial intelligence and how it is developed, perhaps we may finally begin to see a turn in the tide. 

Rachel Varnham

Rachel has an Undergraduate Degree in Sociology and a Master’s Degree in Creative Writing. She enjoys writing about academic research, specifically focusing on sustainability, environmentalism, and gender equality.

About BlueSky Thinking

Business school thinking rarely escapes the campus. BlueSky Thinking changes that. We explore the research, ideas and expertise of the world’s leading faculty, management thinkers and practitioners – enabling their expertise to meet real-world ambition, alongside the rankings, study options and career insight that help ambitious applicants make smarter decisions.

Interested in this topic? You might also like this…

Leave a Reply