Business leaders in the UK recognize female leadership will be vital for responsible AI implementation, but this is not being reflected in representation in senior roles at firms, research reveals.
A new report from IBM investigated how female leadership could shape AI development moving forward, surveying over 4,000 senior business decision makers, split equally between men and women.
IBM’s results show business leaders understand the value of gender equity, with 69% of UK-based business leaders agreeing that it’s important that female leaders are involved in AI decision-making.
In addition, 73% of the respondents said they believe improving female representation in senior positions will help mitigate AI-related gender bias issues.
A more gender diverse senior team will also deliver positive economic impact, according to the report, with 74% of respondents stating they see increased female leadership as important to ensuring that the economic benefits of AI are equally felt across society.
UK firms are failing on female representation
But the actions of UK business leaders do not reflect this belief, however, with just 37% of respondents reporting that advancing women into leadership roles is a top priority within their organization.
This is the lowest figure of all the markets surveyed in Europe, the Middle East, and Africa, where on average 51% of respondents said their company was prioritizing advancing women into senior positions.
An equally concerning statistic highlighted by IBM is that 20% of UK respondents suggested a lack of interest in diversity and inclusion from tech companies was inhibiting more women from entering leadership roles.
Commenting on the results, Dr. Nicola Hodson, chief executive at IBM UK & Ireland, said boosting female representation is not an exercise of meeting diversity quotas, and that although UK business leaders recognize this, there is still some catching up to do.
“Ensuring female leaders have a seat at the table in the age of AI is not about ticking a diversity box, it is a strategic imperative,” she said.
“Clearly, UK business leaders recognise the importance of this. But progress is still needed to equip women with the skills and confidence to position themselves at the forefront of this revolution - steering it towards a future that is inclusive, ethical, and enabling for all.”
“Taming AI will be instrumental in making AI safer and more reliable”
IBM’s study noted that women typically exhibit a better awareness of the ethical considerations concerning AI deployment. When asked to identify the most important qualities of a business leader in the age of AI, the most popular answer among female respondents in the UK was ‘knowing which roles can offer the best strategic advice’.
For their male counterparts, on the other hand, the most critical qualities of a business leader in the AI era are their technical expertise and the strength of their understanding of the AI supplier landscape.
Speaking to ITPro, Eleanor Watson, IEEE member, AI ethics engineer, and AI faculty at Singularity University, said women are particularly well-placed to help address ethical issues in AI systems.
“Women are stereotypically skilled in empathy, 'tending and befriending' others, and carefully molding children. These natural talents can be applied to shaping the values of increasingly interpersonally aware AIs, teaching them to respect boundaries and act in culturally appropriate ways, whilst learning individual preference”, she explained.
“These processes of taming AI will be instrumental in making AI safer and more reliable, enabling it to fit into our personal and professional lives with full flourish.”
Commenting on the problem of various forms of bias in AI systems, Watson noted that this can be addressed according to how the system will be used, and as such organizations need to be careful to avoid adopting a one-size-fits-all approach to implementing AI tools.
“The extent of this bias appears to vary depending on the context in which the model is used. For example, when generating text related to professional scenarios, such as querying about top jobs for men and women, the model's outputs may show less pronounced bias”, Watson said.
“However, when it comes to generating fictional narratives or stories, the model may exhibit significantly more gender bias, as evidenced by diverging character portrayals based on gender. The variation in bias across different contexts implies that even if LLMs might appear unbiased or neutral in some situations, they may still perpetuate harmful stereotypes in others.”
