The clock ticks 6:30, and our big day finally arrives. So we march on, passing along the wooden bridges, stone houses, arid hills, and snow-capped mountains. 

In a bit, I am at the front, presenting our work, and teaching how an AI model works. But I was not merely teaching. I was learning too, learning about my journey of Nesta Fellowship that had started a few months ago. 

I still remember the day I found out about the Nesta Data Science Fellowship, kindly funded by the UK Humanitarian Innovation Hub (UKHIH) as part of the Collective Crisis Intelligence project. Coming back from a sabbatical, I was excited to come across an opportunity where I could use Data Science for a social good project. So it started with plenty of eagerness and enthusiasm, getting to know new people and about the project, and of course going back to roots (Python and Jupyter notebook), which reminds me of our Data Science team.

Working with Izzy has been such a fun learning experience. Be it technical skills (like structuring Machine Learning pipelines) or work management skills (like dividing and collaborating on Data Science projects), I have learned valuable Data Science skills from Izzy. Then there is George (at the beginning) and Jyl (towards the end) from whom I have learned important decision-making skills (like seeing a Machine Learning model more from a usability point of view than just algorithms used, or making big pivot decisions like ‘Regression to Classification’ when needed.) Through regular updates, feedback, and iterations, I got the perfect opportunity to improve my Machine Learning skills in a typical Machine Learning manner. 

Beyond Data Science and tech development, I also experienced one novel feeling right from my beginning days. It was the feeling of transitioning from the usual “how to find” to “what to find” and “why to find”. Looking back, those initial involvements in key planning and decision-making processes made me feel more part of the project, and definitely more accountable. 

So Data Science, tech development, and decision-making processes, all facets covered. What else? 

Of course, there is one more facet left. And that’s Collective Intelligence. 

Whenever I think of Collective Intelligence, I think of activities. And whenever I think of activities, I think of Aleks. 

To be totally honest here, initially, I didn’t see the community engagement activities as something as important as the Machine Learning model. Therefore, it was solely the passion and dedication of Aleks — be it her weekend-prepared detailed presentations, or her meticulous framing of activity questions (not to mention random hour google doc comments) — that made me think, “Okay, this seems serious!”

And yet, the transition from “serious” to “seriously important” required more. It required the fuel of my direct involvement, which I luckily got to have in community engagement workshops.

Compared to my earlier works where the end-users were either imaginary or not even figments of imagination, interacting with the locals made me see our work in a different light. It not only made me understand what was at stake (thereby making me even more accountable and cautious about our model) but also taught me how things could go wrong despite the good intention (like knowing about the reluctance of locals in using Ethnicity which we could have never captured through our model.) 

Rather than seeing participatory workshops and human feedback as ‘compromises’ in the AI model, I started seeing them as necessary ‘complements’. This is where I realized how Collective Intelligence could fill in the gaps of AI. 

Then came the evaluation workshop, the backdrop of this story. 

Already buoyed by the enthusiasm and curiosity of local people, seeing Kathy and Aleks work every bit with them (despite the language barrier) taught me one important lesson — that there are myriad ways one can contribute to any project. 

Then came the moment, my presentation upfront. 

With every explanation I was making, I was also learning more and more about the things I had learned. By the time I finished my presentation, I had realized it was my own evaluation too, the evaluation of the learnings and the transformations I have had because of my work in Collective Intelligence. 

Of course, Data Science (and programming) is still my primary task, as I was reminded on my very first day of homecoming (in Jupyter notebook, where else?). But my experience as a Nesta fellow has helped me think of AI in terms of Collective Intelligence, and see human faces beyond each data point.