With the Yin-Gemis project we want to draw attention to the experiences and stories of women working in the tech industry. We collect stories where women face prejudice because they have IT/technical related jobs. This allows us to research and visualize a reflection of the current industry. We created an AI-generated story, using AI to identify common themes and experiences and created a story, with the aim of drawing attention to the gender gap.
This story was created with AI, through the analysis of the transcriptions of 22 interviews and visualized with VR(web). You’ll be able to see the questions we asked, as well as listen to the AI’s response.
We presented our project to Data-driven design Master students at the HU in Utrecht on November 1st. The guest lecture was given as part of the Philosophy of a Digital Society course. We presented our concept, method, and the AI-generated text (with Audio this time) and a VR sketch we created to give an impression of our next step. We also had a discussion on the topic and gave the students the opportunity to speak with us about it. We were really interested in hearing their thoughts on the gender gap in technology. The students were highly engaged and were very critical. It was mostly about how and why the Gender Gap existed in the first place, as well as the awareness and consciousness it has on design and society. Some of the students recognized some of the insights from the interviews. In addition, a couple of them came up with some design ideas, such as building a platform that looks at talents rather than names or gender. Another student suggested that through establishing accountability, company boards, or those in charge of management, could be more involved in providing a good work environment for all employees.
We appreciate the opportunity to talk about our project and have a fantastic conversation on how to utilize design solutions to overcome the gender gap in tech. We would like to thank Dr Dennis Nguyen for hosting the guest lecture!
Books in the background: De IT-Girl (The countless examples and convincing figures on the Gender Gap in Tech) by Chantal Schinkels & Invisible Women: Exposing Data Bias in a World Designed for Men by Caroline Criado Perez
So far, we’ve interviewed 22 people and transcribed and analyzed the data from 19 of them. Because the libraries (machine learning models) perform best on English material, we translated the Dutch interviews into English. The transcripts were then converted into the format required by a machine learning open-source text generator that we discovered. This text-generator uses Keras, a Python-based deep learning API built on top of TensorFlow’s machine learning framework. We also anonymised the data by assigning a unique code to each statement and categorizing the sentences by question. This makes it easier to see how the sentence (response) came to be in the first place.
We ran a light pre-analysis before executing the code, looking at the quotes and categorizing them into insights, and looking for a relationship between the data. The father as a role model and influencer in career choices, overall bias and prejudice on family planning, prejudice concerning underestimating, participants’ opinion on a quota, working in a diverse team, and message to young women were among the themes that aligned. Overall, it appears that a young woman’s job path is heavily influenced by her father. The father ‘opened their eyes,’ introduced technology into their lives, or encouraged them to seek a profession in technology because they believe it is a good fit for their skills. We also see that the majority of women encounter various forms of bias, whether it is related to family planning, remarks about whether or not they plan to have children, or how their work would be affected. Another trend we noticed was underestimating. The majority were either underestimated and had to prove themselves capable first or if they were capable at all. Working in a diverse environment, according to the majority, is the ‘best’ since it changes the ‘dynamic,’ ‘atmosphere,’ and ‘culture,’ as well as adding a ‘creative approach.’ When we asked the women what they thought of the quota, the general reaction was critical, mostly because it makes you a “diversity hire,” that people should be “hired based on their skills,” and that it is an “artificial thing” to impose. Companies, on the other hand, require that push to ‘break some things’ and ‘reset’. ‘People tend to hire people who look like them,’ said a few, and the quota changes this. Finally, the majority of women urge other women to pursue careers in tech and to “just do it!”, the industry is diverse and offers many opportunities. We’ve also heard that women prefer working with men and that doing so is ‘fun.’
We next ran the algorithm, training the model on our data and asking it the same questions we asked the participants. The model’s generated text appeared to us to be quite familiar at first glance. We chose a few quotes from the generated text. We didn’t change any words or the structure of a sentence in any way. Only punctuation marks were added to a sentence. Eventually, we will compile all of the quotes into a single story and present it in a virtual reality (web) setting.
Meet-up Creative Coding Utrecht
We presented our work in progress to makers, educators, and people working in tech during a Creative Coding Utrecht meet-up, along with our process and insights so far. Because the project is still in progress, input helps the project evolve and be enriched by other people’s perspectives and critical look.
Overall, our project was well-received, with people appreciating the issue and its relevance. We received some useful and insightful input. To begin, we must be aware of implicit bias and add nuance to the work. We explained that the tech industry lacks femininity since there is a lack of women (15.6 per cent women in tech, in the Netherland, CBS, 2021). Gender is not black and/or white, according to the feedback we got, and by introducing nuance to our work, this could help. Another opinion was that not all workplaces where men make up the majority of the employees are unpleasant and that certain workplaces are lovely to work in, and that male collages are pleasant to work with. And the ‘worry’ was that it wouldn’t appear so with our project. Another comment, on our observation that a father plays a vital part in a daughter’s life; was also acknowledged during the meetup. We received comments and questions concerning the end product’s presentation; it was unclear how we would convey the final AI-generated story. What’s the connection between AI and virtual reality? And who is the intended audience for the final product, as well as whether VR is sufficiently accessible? The following stage caught people’s interest in general.
Several things were evident to us as a result of the meet-input. We need to clarify a few things, primarily our motives and methodology, as well as give the project additional complexity. Let’s unpack the above. With this project, we want to showcase women’s perspectives on working in the tech industry by merging all the interviews into one AI-generated narrative to create awareness of the lack of women/femininity in tech. We understand that gender is not black and white, that femininity and masculinity aren’t either. For our project, we decided to choose a focus, and our focus was on individuals who identify as female and have work experience in tech. By bringing more nuance into the project, and by adding quotes from the interviews, to the AI-generated text, we will show that there is a different meaning to the output, where it came from and the many ways it can be interpreted.
We use AI mostly because we want to convey the story in a unique and creative way. It also demonstrates the women’s stories’ alignment, and it would become an anonymous story that does not actually belong to anyone. It’s a gathering that allows for personal interpretation. Furthermore, reading that story will need less effort, and it will hopefully create greater awareness, as opposed to reading a report or paper, as it may reach a different audience. To convey this story, VR will provide an immersive experience, also by using audio and visual components, and it will then leave a stronger impression. We’ll use webVR because it’s more user-friendly. The Louvre Online Tour, for example, allows users to digitally stroll around the Louvre and explore its masterpieces. Viewers do not need VR glasses to explore the area; all they need is a browser. Of course, VR glasses will enhance the experience, but they are not required.
Lastly, this AI and VR story is for individuals who work in tech, so that they may relate to it. We also seek to raise gender awareness among boards, companies, governments, and other entities by focusing on the gender gap and the resulting consciousnesses.
We’ll next generate more AI text and ‘write’ the story, visualize it in a virtual reality environment, and test it before releasing it. In addition, we will perform a thorough analysis of the data, conduct more background studies on femininity and masculinity by consulting various sources and speaking with experts in the field.
We’d like to thank Creative Coding Utrecht for hosting the meet-up, as well as everyone who attended, provided constructive criticism, and, most importantly, all of the participants we’ve spoken with thus far.
The project ends in November/December 2021.
Curious about our project or do you have questions/remarks or do you know a woman in Tech who might be interested in talking to us? Feel free to contact us at julia @ yingemis.nl or rhied @ yingemis.nl
The research is non-commercial and subsidized by the Stimuleringsfonds, Digital Culture.
We spoke with women (22) in the tech business from all sorts of backgrounds, e.g. development, architecture and engineering, consultation and management. Conducting additional interviews adds to the project’s value and insights. Are you interested in talking to us and sharing your story and experiences? Reach out by contacting us via julia @ yingemis.nl or rhied @ yingemis.nl. The interview lasts approximately 30 minutes to a maximum of 45 minutes.
We record the audio conversation session, but make the data completely anonymous. Alternatively, we can also conduct the interview via email. We also send the transcripts for final approval before they are used for the project. More concretely, the anonymized interview data and anonymized key quotes would be used.
The research is non-commercial and we would like to share all results and practical results (eg advice and tools) with you. This project is subsidized by the Stimuleringsfonds, Digital Culture.
Photo: Matt Botsford