Finding data outside of my own reality bubble
I previously wrote about how our environments, biology and interests shape our understanding of the world. I also discussed how this understanding makes us oblivious to certain data needs for fair and safe AI systems.
This led me to wonder: how does my understanding of the world make me oblivious to certain data needs? And, more importantly, how can I become more conscious of these data needs?
Starting my own experiments
To answer the above two questions, I’m starting some of my own experiments. In these experiments, my goal is to move outside of my own reality bubble in order to encounter data I am unaware of.
I’m not quite sure yet what data I’ll explore exactly, but let’s figure that out along the way. The important part of the experiments is to get outside of my own reality bubble.
People-centered
Getting outside of my own reality bubble is not something I can do on my own. Within the field of AI, getting in touch with other people’s perspectives – either through colleagues or end-users – is essential. I will follow this tradition in my experiments.
Ultimately, my aim is to learn about other people’s realities. Obviously, I can never truly leave my own reality bubble behind.
Nonetheless, I hope my experiments give a voice to the people who’s data would otherwise not have been considered when building AI systems.
My first experiment will be Email Hitchhiking.