24 September 2026
City of Glasgow College

The industry roundtable events have become a firm fixture of the Ada Scotland Festival.  In this session, festival partner Royal London led discussion around the question:  As AI becomes increasingly influential in shaping business and society, how do we ensure women are helping to shape the technology, rather than simply adapting to it?

Opening Remarks & Context

Seatbelts were designed for average size males, and this design bias means females are at higher risk of injury in traffic accidents. 

When just 14% of tech leaders working in AI are female, the technology is designed primarily by men for men which can create design bias – can this be mitigated?

The Workday Scandal in the USA illustrates the potential impact of bias inherent in AI models. The high-profile, class-action lawsuit alleges the AI model used by Workday discriminated against candidates based on age, race, and disability.

This raises questions around how to control bias in building generative AI solutions.  Are we experiencing the “seatbelt phenomenon”?

Generative AI is being used to create “efficiencies” in organisations, already leading to redundancies.  Does AI’s continued growth undermine people’s access to employment and dignity in work?  Are we training or building technologies with the potential to replace us?  When is more not actually more?

What are the issues?

• Importance of an intersectional approach to filtering out bias by gender, race, sexuality, disability etc.  Generally, greater awareness of biases in the public than private sector.
• Representation is the solution, but the cause is the structural inequalities which cause under-representation in the first place.  This is evidenced in every part of our lives – from the statues in our built environment to the seatbelts we wear and AI is trained on data which reflects these biases.
• Organisations with capacity to test their own models are better able to calibrate for diversity as they have control over the input. However, the vast majority of organisations are implementing “off the shelf” models which makes it difficult to realise ethical or responsible AI obligations.
• Fintech companies are typically not innovating with AI – instead they are buying black boxes and considering how these can create efficiencies
• Are we buying tech we cannot control?  We have no control over the construction of the models including their political shaping.  Humans need to harness and control AI, and not the other way around.
• You can hire or work with people who hold unacceptable views provided those views provided they get the job done.  Is this applicable to AI?
• Government regulation is required where output is monitored to reduce biases.
• Access to AI is not the same as influence over AI. 
• Females are underrepresented in tech and STEM and have higher dropout rates at every stage with multiple causes, but generally females are interested in tech and we have a responsibility to ensure educational and workplace cultures facilitate inclusion.  

Concerns

• There is a risk people trust information produced by generative AI as a source of “truth” without questioning the output.  Where the information is produced with bias, this can mean users become inadvertently biased themselves.
• AI is most effective when the user has the knowledge to critique the outputs.  However, AI can undermine the practice of critical thinking skills.  This is a particular concern for young people who are still developing critical thinking and ethics skills.
• There are concerns the workforce of the future will not have the opportunity to develop important skills because of a reliance on AI.
• Challenge for educators in how to adapt curriculums to account for AI across all disciplines: should we teach ethics instead of AI?
• The people most at risk of redundancies because of AI adoption and implementation are women as “feminised” roles are the first being replaced with AI.
• Men are praised for using AI in the workplace, but women using it are perceived as less able.
• Loss of creativity and challenges for creative designers specifically
• High level of misinformation around AI due to a lack of trustworthy sources.  

Positives

• Green shoots visible in gender diversity improving in STEM education and early careers in tech and at more senior levels in some organisations.
• EU AI Act: first of its kind legislation which is a positive step in setting a precedent for regulation and governance. 
• The Act requires entry level AI literacy for whole populations routed in uses, rather than technology.  The aligns with an understanding that AI is not a tech problem but actually about literacy, structural inequalities, representation, leadership, sustainability, resource extraction etc.
• AI can be used to help resolve some biases e.g. better understanding the gender equality gap in pensions by mining data using AI; medical advancements
• AI cannot shape our future, we need to shape the future of AI
• It is creating new roles in the workplace
• Higher quality inputs can create higher quality outputs
• AI use can help navigate busy lives and boost productivity
• Some AI models are fun and there is enjoyment in engagement with them

Concluding Thoughts

• AI represents the biggest tech change we will see in our lives
• There are many tensions at once: we can enjoy using AI and making efficiency gains in our lives while also being aware that AI can compound inequalities, reduce access to society and transform the workplace, replacing many roles.
• Importance of remaining a curious user involved with the topic and keeping societal lenses in mind.

Thank you to Ciara Conway for expertly chairing the discussion, to Royal London for sponsoring, City of Glasgow College for hosting and to all who attended and shared insightful contributions on a challenging and important topic.