Workshop preparation & AI expert Interviews
Final Major Project blog 5
Week Commencing 23 September & 30 September
Workshop planning
We planned our research methods for the workshops with a focus on fostering connections between people through storytelling. To achieve this, we chose the free flow writing technique, where participants write freely without worrying about grammar or spelling (Alharthi, 2021). We planned for this to lead into a group reflective discussion. Followed by the creative toolkit method.
This method is a collection of physical elements for modelling, visualisation, and creative play to inspire design (Martin and Hanington, 2019 ). Using this approach, we aimed to physically reimagine AI to support the experiences discussed in free flow writing.
Finally, we prepared a presentation, agenda, and a trial run of the experience for feedback. I then met with the Post-Grad community , taking them through our presentation and created a proposed budget plan to get support with purchasing materials and refreshments on the day of the event.

Fig. 1 Extract from my free flow writing and our outcomes from testing the creative toolkit activity
Speed dating ideas with AI Expert
We had two meetings with an AI expert and Trans Women who shared their experience of learning about and designing with AI. They also feedback on our Storyboards and explained AI datasets to us, helping us realise AI is not as sophisticated as we had assumed.
They explained AI is talked up in the media and as a result is seen as many things. Furthermore, they helpfully suggested we focus on specific aspects of the trans experience, noting that our exploration related more to their transitioning phase than their current post-transition life.

Fig. 2 Interview aims and findings from interview with AI expert
Interview with an AI expert
Teddi also interviewed an AI expert who suggests Generative AI is a glorified search engine which is mostly changing the way people consume information. They suggest we could potential use the deep scraping element of AI to benefit the Trans community. For example, through prompt engineering people might be able to access information which is difficult to find on the internet.
This links back to an earlier interview with Trans in the City who suggested Trans information is difficult to find online. Prompt engineering is giving a large language model a set of instructions or a question in human language to elicit a particular response (Dwivedi et al, 2023). Teddi and I would like to use this technique to explore information the LLM provide about the trans community.
Reading: What's the buzz with AI
The Data detox kit by Tactile Tech (Ghnaim, 2024) categorised AI into three categories from image generation, text generation and recommendation systems. A reminder it might be helpful for Teddi and I to narrow down to a specific AI. To do this I suggested we map out our research against the three categories to understand what we have covered and where we want to focus.
Currently we are interested in algorithms due to their influence on trans peoples everyday experiences either overtly or covertly in the case of the algorithm scanner at airports in Constanza-Chocks reading Matrix of Domination (2018).

Fig. 3 overlaying our research onto the different types of AI explained by Tactile Tech (Ghnaim, 2024).
Trans-forming AI Workshop
Final Major Project blog 6
Week Commencing 7 October
Workshop testing
Before testing we added step-by-step activity explanations with timings for clarity and included a feedback form to gather participants' perspectives and suggestions for improvement.
Saranya and Srushti tested our workshop and provided feedback on its flow, content, and language. For example, they found terms like gender expression and identity unclear during the free-flow writing activity, so Teddi added a clarifying slide. They also suggested removing some early slides explaining AI. To improve engagement, they proposed moving the free-flow writing activity to the beginning and moving our research slides to before the creative toolkit activity. Teddi rearranged the slides accordingly, resulting in a much-improved output.
Reflections: Teddi and I discussed the free-flow writing prompt in detail, we wanted to explore the evolving journey of being trans+, including and beyond 'transitioning,' as gender expression is constantly evolving. Testing the prompt with a trans+ person might have highlighted the community's familiarity with these terms.




Fig. 4 Our final slides for our Trans-forming AI workshops
UAL Post Grad Community Trans-forming AI Workshop
We initially planned to use two tables for the workshop but had to adjust upon arrival. We opted to use one table for activities and a sofa area for presentations. This demonstrated the difficulty with planning all aspect of a workshop and how flexibility to adapt can lead to better outcomes. With support from the Post-Grad community, we purchased soft, malleable materials for the creative toolkit activity to explore the concept of a more "caring" AI.
During the workshop, participants enjoyed discussing their experiences with others in the Trans+ community while exploring AI in a unique way. Although it was challenging for people to conceptualise physicalising AI, the lively, collaborative atmosphere showed they enjoyed creating together. Teddi and I realised our note-taking needed better planning and we will improve this next time. Thankfully, Alejandra and Sean came to capture photos and videos, greatly aiding our documentation.

Fig. 5 Images from the workshop including participants doing the free flow writing (1), sharing what they wrote (2), Teddi and I going through our research (3), participants doing the creative toolkit (4) and participants sharing what they made (5&6).




Fig. 6 Images of peoples creations during the creative toolkit activity and quotes from their explanations




Fig. 7 Feedback from those who attended and filled in the feedback form.
Presentation and Tutorial
During our progress presentation, participants wanted to understand our design outcome: are we producing an AI, critiquing an AI, developing a method, or creating an approach to using AI? They also raised concerns about us relying on verbal discussions about consent and ethics in the workshop instead of signed forms. In response Teddi created an ethics form for the next workshop.
They appreciated the sharing of experiences and stories and questioned whether we were advocating a return to peer-to-peer (non-algorithmic) interaction. While they valued our focus on care, Alaistair cautioned against anthropomorphising AI and Tonicha noted if AI currently excludes the trans community, it may not be empowering to develop AI for it.




Fig. 8 Presentation updating our project to class


Fig. 9 Ethics and consent form for workshops
Reading: Atlas of AI
One challenge during our workshop was conveying the purpose of physicalising AI in the creative toolkit. Reading the introduction to Kat Crawfords Atlas of AI (2021) clarified our reasoning, as it explains AI as embodied and material, rooted in natural resources and human labor, rather than truly artificial or intelligent. This aligned with our goals for the creative toolkit. Furthermore, during the workshop, participants created items like flowers to symbolise their desire for the ecological impact of AI to be considered in it’s creation.
References
Alharthi, S., 2021. From instructed writing to free-writing: A study of EFL learners. SAGE Open, 11(1), p.21582440211007112.
Costanza-Chock, S., 2018. Design justice, AI, and escape from the matrix of domination. Journal of Design and Science, 3(5), pp.1-14.
Crawford, K. (2021) Atlas of AI. London: Yale University Press.
Dwivedi, S., Ghosh, S. and Dwivedi, S., 2023. Breaking the bias: Gender fairness in LLMs using prompt engineering and in-context learning. Rupkatha Journal on Interdisciplinary Studies in Humanities, 15(4).
Martin, B. And Hanington, B. (2019) Universal methods of design. Beverly: Rockport Publishers.
Ghnaim, S. (2024) 'What's the buss with AI', Data Detox Kit, 23 August. Available at: https://datadetoxkit.org/en/ai/buzz/ (Accessed:13 October 2024).