In celebration of International Women’s Day 2022, a report on ‘The Effects of Artificial Intelligence on the Working Lives of Women’ has been published by the Inter-American Development Bank (IDB), The Organisation for Economic Co-operation and Development (OECD) and the United Nations Educational, Scientific, and Cultural Organization (UNESCO). The report was researched and prepared in acknowledgement of the challenges faced by women around the world in achieving greater participation in STEM employment and parity of earnings across all industries. Within its research, the report considers the ‘hidden’ additional workload of women in terms of unpaid work, such as childcare. The report recognises the potential of AI technologies to support women while also highlighting the risk of AI exacerbating the current systemic gaps between working men and women if the risks it identifies are not addressed.

Drawing on the contributors’ knowledge of the effect that AI technologies can have on women’s potential to gain employment, and how employed women are treated in different regions of the world, the report outlines six key recommendations. The report does not claim to be conclusive, acknowledging that the ways in which AI is applied to the working world will undoubtedly change over time, but rather emphasises the need for continued research and exploration of this area.

The report’s six key findings are:

  1. Reskilling and upskilling women workers – AI is changing the labour market and the skills expected from workers. It is therefore crucial that women are not left out of the increasing demand for STEM/AI related skills.
  2. Encouraging women in STEM – More women are needed at the forefront of AI than ever before. To enable this to happen, steps must be taken to support women towards STEM/AI focused career paths.
  3. Accounting for contextual and cultural complexity – The impact of AI will always depend on the context of where and how it is being implemented. Contextual factors, such as cultural complexities, must be accounted for when designing and implementing AI systems.
  4. Leveraging multi-stakeholder approaches – It is not enough for private companies to be at the forefront of involving women in STEM/AI focused careers. Governments must equally ensure, through skill-development initiatives and similar programmes, that there is a skill-equalising environment available to women.
  5. Shaping gender stereotypes – The report shows the substantial impact that stereotypes with respect to the paid and unpaid work of women, and how both can be shaped and accounted for by AI. Virtual personal assistants, and similar software to which gender is applied, need to be appropriately considered during the design phase to prevent the continuation of outdated stereotypes.
  6. Continuing applied research – As with all technologies, continued research is a necessity; particularly with regards to how (and to what extent) this impacts women. By understanding these impacts, we will be able to bridge some of the gaps through a tailored approach to the implementation of AI both in and outside of places of employment.

It must be noted that many of the findings of the report could apply equally to those identifying anywhere on the spectrum of non-male gender or as genderless. It is of crucial importance to bear in mind that the issue of potential unconscious bias through the application of AI applies to a wide range of human characteristics. In Session 2 of DLA Piper’s current webinar series looking at AI in the Workplace, our panel discussed the key risks posed by utilising AI in the workplace, and recognised the complex journey ahead in removing unconscious bias from AI systems being used, for example, to review job applications. As Jennifer Holyoake, a Legal Director in our Employment team, noted during the session, AI is not neutral because it is driven by human choices, such as the objectives of AI and the data selected to train the algorithm. This means that the outputs are not impartial and there is a risk of perpetuating existing bias and, furthermore, of this leading to systemic bias. This echoes the concerns of the report of IDB, OECD, and UNESCO, and emphasises the very real need for both the continued technical development of AI systems to minimise bias and, crucially, the additional development of risk and impact assessments, controls, and regulation to ensure that AI is applied fairly and in equal support of all humans, regardless of their characteristics.

The next and final session of DLA Piper’s AI in the Workplace series will go on to consider a framework for AI success in the workplace and will be webcast on 16 March 2022. To register for the event or to listen to the recordings of any of the first three sessions, please click here.

For more information on DLA Piper’s AI in the Workplace series, or if you have any questions relating to this article, please contact the authors or your usual DLA Piper contact.