When we talk about the future of the workplace, technology – and AI in particular – is inevitably a key part of the conversation. A 2023 study found that 37% of people in the UK have now used AI at work, so it’s important to question whether it is being implemented in a way that isn’t perpetuating existing biases in the workplace.

Research by the University of Melbourne looking into whether AI will empower or disempower women in the workplace found mixed results. “It may be cementing gender biases, but on the other it could be creating safer workplaces,” the report reads. There seems to be good and bad that comes with the increasing use of AI in the workplace, and the key is using the positives to neutralise the negatives.

For example, when implemented mindfully, AI can contribute to making workplaces safer. The technology is being used to combat sexual harassment, examples including Callisto – which enables sexual assault survivors to report securely and to connect with survivors of serial offenders. AllVoices, is another AI tool that can be used to report sexual harassment anonymously. In this case, the collation of data can also be used by the company to determine the scale of the problem at large.

Positive psychology and business coach Elle Mace says that these projects prove that AI can “provide a more systematic and efficient approach compared to human-based interactions” when it comes to sensitive subjects like sexual assault. “It allows for quicker detection and response, providing support to victims in a timely manner,” she adds.

Cecily Motley, co-founder of Harriet, an AI-powered workplace assistant, insists that while AI should not be viewed as a replacement for human interaction, it can provide a “vital gateway to that support”. That said, human monitoring of these systems is crucial to ensure that survivors are using a system that will support them appropriately.

“The key thing is that humans closely monitor models to ensure that they are working in an ethical and secure way, and to provide human support and follow up care for those who need it,” Cecily says.

Use of AI during the recruitment process can also reduce gender bias, with a study finding last year that AI is more likely to hire a woman than a human is and the majority (96 per cent) of job adverts at FTSE 100 companies have also been found to use gender-biased language.  

Use of AI could almost double the number of women considered to be in the top 10% of candidates. While the opportunity to empower women through the use of AI is possible in this case, though, the adverse result is also possible if the software is programmed with biased data. 

“AI is driven by logic, taking into account only facts - but it is dependent on the bias that it is pre-programmed to evaluate with,” C-suite business coach Kathleen Anne Harmeston says. “If AI is trained with biased data – for example if we teach it with criteria from human interviewers — it may not just emulate, but also exacerbate the human bias.” This can build on huge endemic examples of gender inequality in the workplace, such as the gender pay gap.

So a huge way that business leaders can empower women through AI is to ensure that human bias isn’t reflected in its programming. “Fortunately this can be managed by teaching AI to predict relevant and objective outcomes, rather than emulate human intuition,” Kathleen adds. It’s all down to how the technology is implemented in the workplace.

“Algorithms can be designed to focus on skills and qualifications rather than subjective judgments that may be influenced by unconscious biases,” Elle agrees.

Of course, another large part of gender bias in AI and its potential for disempowering women in the workplace is the fact that so few women are involved in its development. More than 80% of AI professors are men, while 18% of authors at leading AI conferences are women. If there’s a lack of representation in the production of AI, how can we trust that the technology is devoid of the gender bias we already find in real-life (also often male-dominated) workplaces? 

So as a larger scale solution, there needs to be more encouragement for women to pursue a career in STEM subjects and enter the respective industries. This way, Kathleen says, they can “address the bias which exists in all aspects of this employment arena”, and can contribute to the development of AI as a product, so that they are “designed and programmed to neutralise any pre-existing bias.” 

Elle adds that increasing the number of women working in AI “could bring fresh perspectives, potentially leading to more inclusive and unbiased algorithms”. 

A key example of male bias in AI is the default use of a female voice in personal assistants such as Alexa and Siri. Use of mainstream AI technology in the workplace that automatically infers the female role as one of “assistant” status is arguably allowing a subtle form of gender bias into working culture. 

“Their tones are docile and obliging and may send the signal that women are subservient and reinforce the idea that such roles are for females only,” Kathleen says, adding that the problematic normalisation of inequality and disempowerment of women doesn’t end there. “What is more disturbing is that these devices are often programmed to deflect or lack agency when given insulting commands.  It can be argued that this precedent can reinforce gender biases i.e. that women are subservient and will tolerate poor behaviour.”

When it comes to using these AI tools in the workplace, Elle advises that there are ways to ensure that their usage empower women, including “diversifying voice options and allowing users to choose their preferred voice, which could help mitigate this issue and promote gender equality”.

If implemented correctly, AI has the scope to liberate women from more administrative jobs, which they incidentally are more likely to be doing than men – a 2022 study found that 74% of admin or clerical workers are women. AI can take up these tasks, leaving space for women to progress in more creative or executive areas.

“If women are released from admin-based tasks, AI will allow them to showcase their real skills within a role,” Kathleen says, pointing out that the individuals running a workplace are then responsible for empowering them to progress. “It is then up to the culture of an organisation to identify those skills and allow them to flourish through unbiased promotion opportunities.”

AI’s take-up of admin in the workplace could act as a double-edged sword though, as it could render certain jobs redundant – with women being more vulnerable to this than men. According to a report by McKinsey, more women than men stand to lose their jobs by the end of the decade because of the rise of artificial intelligence and automation.

“Women from diverse backgrounds without education and from disadvantaged segments of society will be at risk as well as the youngest and oldest female workers,” Kathleen says. 

For this reason, one of the most important things that business leaders can do to ensure that their implementation of AI in the workplace empowers women is to use the benefits of this technology to neutralise any gender bias or inequality that it could exacerbate. 

According to Stacey-Rebekka Karlsson, managing director and founder of female-led marketing agency Goho, it is crucial that business owners like herself “involve AI in creating gender-inclusive policies in the workplace, ensuring that decisions are based on data and not unconscious biases”. 

She adds that ongoing checks for biases in AI algorithms and datasets should be implemented, as should ethical guidelines for its development and deployment. AI can also help women to balance caregiving responsibilities with work, as it allows them to automate time-consuming tasks, as well as facilitating remote working.

Businesses would also do well to provide women with more opportunities to learn about and work with AI, in the workplace and outside, to further empower them to meet their career goals and contribute to the technology’s gender inclusive development.

Above all, there needs to be an awareness of the inequalities that technology like AI can exacerbate. Beginning with a lack of women working in STEM, and spanning from gender bias during recruitment to the possibility of AI making female-dominated jobs redundant.  We then need to carry out active strategies to counteract these biases, as these technologies become more commonplace.

In fact, Elle suggests that “ongoing education and awareness about the impact of AI on gender dynamics are essential for fostering a more inclusive workplace” for all.