In this final blog, the Peer Learning Journey facilitators Ethan Gilsdorf, David Kelleher, Khanysa E. Mabyeka, Lucía Mesa Vélez and Marie-Kate Waller share their learning from accompanying AI4D hubs and labs across the two-year PLJ process.
David Kelleher: Facilitators go on journeys, too: Lessons, learning and takeaways
“Why haven’t they written back to me?” “What can I do to move this process forward?” “Why is this not more of a priority?” These are some of the questions that facilitators ask themselves in this complex relationship with participants in the PLJ (Peer Learning Journey).
As facilitators of the Artificial Intelligence for Development Africa (AI4D) PLJ process, we help support participants as they advance a GEI agenda, building learning spaces that truly engage participants. It is this engagement that allows them to create change in their projects and their organizations.
We have chosen not to follow the well-worn and more comfortable path of training programs and checklists. We know that on their own these techniques seldom engage participants in real change. In fact, they can actually generate resistance. Instead, we value the autonomy and good will of participants and build learning spaces in which they make choices as to how they want to advance GEI in their projects and in their organizations.
However, creating spaces that respond to the needs for true engagement and relationship means that there is no textbook response to situations. Instead, facilitators like us build relationships, respond to situations and people, and improvise.
Yes, facilitators take learning journeys too. The following short pieces describe some of our feelings and thoughts through this process.
Marie-Katherine (Kate) Waller: The PLJ method is not perfect
In helping to organize this PLJ process, I learned the power of mentorship as being a critical friend. This involves asking powerful questions like “How have your multiple identities and own socialization shaped your own assumptions, power and position as a researcher and scientist” and “Could you get your engineering and computer science students also doing qualitative research on rural women’s perspectives of AI?’’
I also discovered the need for experimentation. Scientists need to figure things out themselves, including what works and does not, guided by relevant GEI resources along the way.
That said, the PLJ method is not perfect. I think we did not completely satisfy our PLJ participants. Two hour sessions for three days online is hard to build connections with others. This is despite highly interactive sessions. Learning how to mainstream gender and inclusivity may seem easier by gender training or producing a check-the-box guide. After 30 years, however, we know now that these prescribed methods are not enough to bring about systemic, and meaningful change. Scientists and GEI facilitators alike need to question our assumptions at a personal and professional level. GEI principles are not “out there.” We must embody them ourselves internally and in our relationships.
With this approach in mind, implementing GEI in the PLJ enabled the hubs and labs, and the GEI support team, to experience emergent learning and question our preconceived ideas. For example, despite different fields in AI for development, the PLJ participants across hubs/labs shared the significance of consulting with the marginal AI user, whether a rural farmer, youth living with disabilities or on cultural and linguistic diversities to co-create an AI tool. I saw that learning was mutual across ages, genders, race, class, disciplines, geographic locations and internet connectivity. The PLJ served to create space for questioning, experimenting and peer learning. The Day 3 graphic (below) tells of the collective learning and where these journeys may go.
I appreciated observing that hubs and labs adopted and adapted a kind of “PLJ method” themselves in their work. This does not mean that they did not train their sub-grantees to consider gender and inclusion during their AI training. Hubs and labs also created peer learning processes, from encouraging their sub-grantees to create “GEI change teams” and “GEI change projects” to setting up GEI in AI specific, peer to peer learning forums and workshop series.
As facilitators and PLJ participants, we learned that having knowledge and resources on how to integrate GEI is one dimension. You need to be personally motivated, to see its value, to have the will, and the opportunities to experiment. The PLJ allowed for nurturing the need for the head, heart and feet.
Lucía Mesa Vélez: Touching hearts and minds
I joined the team just before starting the PLJ process, without being familiar with the Gender Action Learning (GAL) online methodology or the HoS (Hearing Our Stories) sessions with the teams. The first workshop made clear the challenge that we had in our hands: retaining the attention of participants for three, two-hour sessions while communicating the information and forming the peer-learning process. I think we had various degrees of success in these areas Participation fluctuated a lot, with few of the participants who engaged since the beginning staying until the end.
However, even if they were few in number, I found it rewarding to see how those individuals who did participate were impacted by the process. I looked forward to the PLJ sessions, because the participants raised very interesting questions and described their progress, from treating the issue of bias in their databases to considering community participation in their projects. By the last session, they showed not only command on GEI concepts, but also a sensibility about the importance of considering gender and inclusion in their projects.
The field of development often focuses only on the numbers. But for me, the quantity of participants isn’t as important as the quality — that my work has impact, touching hearts and minds. I was very glad to hear and read the texts of participants in the last sessions and have hope that they will be agents of change in their contexts. The experience also serves us to polish the online methodology for next time I am involved in leading a PLJ process. For instance, we could try to cover less topics but go deeper into them.
GEI can be a hot button topic. In our work, we certainly confront backlash and difficulties when trying to communicate about it, so having willing receptors is always good. With Phase 2 coming around the corner, I’m curious to see how those individuals carry the learnings forward.
Khanysa Maayeka: The unknown no longer frightens me
began my journey as a member of the GEI facilitation team with anxiety and insecurity. I had heard about AI before and read a few articles both about its benefits and harm by duplicating historical injustices and stereotypes. But I considered myself an AI illiterate.
In the previous two years before my engagement with the AI4D project, I had been involved in another project supporting African Science Granting Councils in integrating gender equality interests to their work. Some of these councils also supported STEM projects, which exposed me to debates about gender inequality in the STEM sector.
However, at the beginning of the AI4D project, I felt that my knowledge of AI and technology development was limited and it was not clear how I could be a resource to the hubs I was to support. When I expressed some of my fears to the then advisor of the GEI support team, he shared that expertise in AI was not necessary and that my experience in analyzing systems and deep structures that create and reproduce injustices was sufficient.
I like challenges, so I accepted this one as an opportunity to practice keeping my curiosity alive and learn to feel comfortable with the unknown. I was fortunate to work with people from the education hub, EduAI Hub, who were open to experimenting with a new work approach:no one was required to have all the answers, and asking questions and voicing ideas, even if they sounded crazy, was welcome.
I remember one significant meeting. After our discussion, one of the participants asked, with a little frustration, “Why are we talking about gender now when we should be talking about how we start engaging with the subgrantees”? This opened the door for other conversations and for other members of the team to voice out why considering GEI was important for them.
Many times, we would end our accompaniment meetings with ideas that were completely different from the ones we had when we began, or even with more questions than answers. This sometimes meant that in order to move EduAI Hub activities forward, more energy needed to be spent in understanding a certain reality, certain dynamic or interests of the sub-grantees.
For example, the team planned to hold a virtual town hall to bring together the developers (their sub-grantees) producing systems to support students with visual or hearing impairments and those differently-abled users. During one of our mentorship meetings, the team realized they were unsure how to hold a virtual meeting for visually and hearing impaired students, and took the step of asking the subgrantees how to best accommodate their potential users. The effort to reach out to the users and customize the meeting for them was a resounding success. This was one example of how we often left the meetings energized by the exchange of ideas and experiences.
In the end, I learned that it didn’t matter with which lenses — computer science, programming, law, social science or political science —the different hub members and myself had begun the project. Our aims were similar: to see the development of technology that contributed to reducing suffering and to ending historical inequalities.
I also realized it did not matter that our entry point was gender. We ended up discussing other factors that intersect with gender and reinforce inequalities, such as discriminatory social norms, disability, language politics, access to technology, urban and rural divide, hierarchies in systems of knowledge, etc. In the end, embracing the unknown took me back to a more comfortable ground, of trying to understand systems and to look for solutions to transform the deep rooted structures in those systems that hold gender inequality and social injustice in place.
I am still AI illiterate. But the unknown no longer frightens me because my experience working on the AI4D project forced me to learn to find value in letting go of the desire to be in control of the content, and to welcome my own curiosity and therefore that of others.
David Kelleher: Participants needed to “get” the idea that inclusion matters
As a co-facilitator during the final meeting of the PLJ workshop, I was impressed as I listened to the discussion. Participants talked about their learning process as they realized that gender and inclusion was an important part of their very technical projects. In other words, they recognized there is an important social component to their work.
Participants needed to “get” the idea that inclusion matters, both as a matter of fairness and also as a matter of good research practice. They were then able to embark on a variety of learning paths to actually implement a GEI perspective in their work. These paths were sometimes intensely social, such as finding ways to consult women farmers without alienating their husbands, or finding ways to help sub grantees learn how to integrate GEI into their projects and following up to help and support them when progress lagged. Much of this learning was counter-cultural and fell outside of the experience and training of the researchers. These people took big steps.
Listening to these researchers made the idea of “socio-technical” come alive for me. Each step of the AI development process has both social and technical aspects to it. Building a team, setting objectives, consulting with stakeholders, building a model and testing are both technical and social activities. Moreover, all of these activities are done within cultural contexts that are charged with collectively held ideas about gender and inclusion. Our job is to help researchers understand, in their own terms, the social dynamics at each stage of their process and help them see how apparently “technical” decisions can have strong gender and inclusion implications.
The Day 3 graphic shares the learning from the last day of our PLJ workshop held in May, 2024. We explored the“ Future: Where are you going?” Participants reflected on what they planned to take forward from their learning about GEI in AI4D research, and what that process might be.
Ethan Gilsdorf: Permission to be a writer
I firmly believe that each person has a story to tell. One simply needs encouragement to reflect, on the page, the arc of their journey.
As a creative writing teacher, I am accustomed to teaching people who have aspirations to be “writers” with a capital “W.” They dream of publishing their novels, stories, poems and memoirs. I also teach what I call “occasional” or “accidental” writers—those who must write white papers, emails, case studies, reports as part of their work, but don’t see themselves as expressive or creative writers.
I would put those who took part in the “Artificial Intelligence for Development Africa” (AI4D) PLJ “writeshop” in this latter category. Yet for these writers, words are just as important, and can be equally transformational.
Prior to my work with this PLJ, I have been fortunate to lead several in-person PLJ and GAL (Gender Action Learning) writeshops, including traveling to Guatemala in 2019 and to Nairobi in 2023. Teaching in person gives me the advantage of seeing people face-to-face, and bringing my energy and force of personality into the room, and making chit-chat and encouraging comments during a coffee break.
This time, the PLJ process for the AI4D project took place remotely. As a writing coach leading sessions via Zoom, I knew my usual tricks and in-person magic would be harder to conjure across the distance of the Internet. I worried I wouldn’t be able to connect on a human level with the participants.
Yet, even if my smiles, bad jokes and kind words didn’t always translate on Zoom, I used the same methods. I provided structure, a deadline, accountability, and some prompts. I began the writeshop with an exercise to get the creative juices flowing: “Think of a significant time when you learned about gender roles from your upbringing, schooling or workplace.” This usually works, and this time around was no exception. In 20 minutes, the participants produced drafts of insightful, poignant personal narratives.
The AI4D participants were then asked to run with that writing mojo to ponder what they learned during their “change project” integrating GEI into AI and their work. They were pushed to articulate what was significant in their journey, or what surprised them, or what was successful (and, perhaps more revealing, what was not). Sometimes I asked, “Tell me something you understood at the end of the project that you didn’t know at the start.” Other times, my co-facilitators or I challenged them to reflect on what they discovered about their team, other teams, themselves, and how their thinking changed. They also were asked to share any takeaways or advice — what to do, or what not to do? — with any stakeholders in the future.
I was thrilled to watch as the writers articulated their process and reflections in sharp, vivid language: “I cracked my head around these questions” “This was my ‘aha’ moment!”; “How to speak out in a place where diversity is squashed?”; and “we, as researchers, had been operating in a bubble.”
I try to be flexible and open, and meet every writer where they are. I respond with curiosity and gentle prodding: “Here is where I became interested”; “I’m curious to know more here”; “Can you provide more details?”
Each person has a story to tell. It is through crafting a narrative — finding a beginning, a middle and an end — that writers come to understand themselves and what they learned. Sometimes all you need is permission to be a writer, and a safe space where writing can be shared. That can happen on Zoom or in person. The writing emerges, every time.
This blog post was written by the Peer Learning Journey writing coach, Ethan Gilsdorf, and Gender at Work and Ladysmith PLJ facilitators, David Kelleher, Khanysa E. Mabyeka, Marie-Kate Waller and Lucía Mesa Vélez. This work is openly licensed via CC BY 4.0.
Curious to read the reflections on gender inequality, exclusions, & AI from this series? Read the other blog posts here.