The impact of artificial intelligence (AI) on society is undeniable. From new life-enhancing medical advancements and increased workforce productivity to huge job losses and stark lack of legal regulation, it holds immense influence, for better or worse. However, it is becoming increasingly clear that AI is having a particularly strong influence on the gender gap, especially in the workplace.

AI is trained using data sourced by its creators and that data can define the system’s biases - while AI itself does not have a sense of morality or a code of ethics, it can inherit biases found in the data. This means, for example, that generative AI systems often exhibit biases in their associations: “doctor” might be associated with “man” and “nurse” with “woman”, due to gender biases found in the data used to train it and the way in which words are encoded to reflect their historical associations with other words. Steps can be taken to minimise this effect, but we are certainly not starting from a blank canvas.

While this obviously has a negative impact – with algorithmic biases perpetuating those already present in society, leading to a particularly potent feedback loop - the problem extends further still. Organisations that use AI to determine employee pay may find that, due to historical gender pay gaps in the workplace, without intervention, the AI creates a gender pay gap to mimic the data with which it has been trained. Similarly, organisations that are integrating AI into their HR and hiring departments, such as in CV screening, may find that it exhibits gender biases based on previous hiring trends, leaving the potential for a biased hiring process. Until legal safeguards catch up, AI and algorithmic biases can have a very real influence on the gender pay gap and workforce demographics.

Part of the problem may stem from the fact that AI is still mostly being developed by men. According to the Global Gender Gap Report of 2023, only 30% of people working in AI were women - just four percentage points higher than in 2016 despite a sixfold increase in talent availability. Similarly, a UNESCO report from earlier this year pointed out that only 18% of authors at leading AI conferences are women, while more than 80% of AI professors are men.

On a wider scale, a report by the Institute for Public Policy Research (IPPR) has determined that women are 40% more likely to have their work replaced by AI in the near future, with traditionally female-dominated roles such as personal assistant and admin assistant expected to be disproportionately affected compared to traditionally male roles. However, given the increasing presence of AI in the workplace, female business leaders, who have consistently demonstrated better positive communication and higher levels of empathy towards their employees, may find themselves at an advantage as they possess skills that cannot be replaced by AI.

Ensuring diversity in the creation process of AI is key to its beneficial integration into society, but there need to be targeted measures in place to protect female-dominated jobs, regulate how AI is used in the hiring process and ultimately prevent AI from widening the gender gap in the workplace.

Elouisa Crichton is a Partner at Dentons

Agenda is a column for outside contributors. Contact: agenda@theherald.co.uk