[Wecount] We Count Planned Co-design clarifications
jtreviranus at ocadu.ca
Tue Apr 13 17:59:19 UTC 2021
Dear We Count development team,
I hope to meet with you soon to help chart direction.
I met with Julia Stoyanovich and we have agreed to collaborate. She is doing similar research and has created great comic books and courses on the topic:
I also want to clarify why we embarked on the mapping exemplar. The map was simply a means of exploring the opportunities and barriers inherent in data sciences for people with disabilities that is relevant and useful during the pandemic. Through the scenario we can identify issues related to community driven data science, minority data, data representation and analysis. Mapping was one means of presenting the data.
The map is not the primary focus of the development effort. We want to develop data science tools and exemplars that serve small minorities and outliers. We also want to create tools that democratize and make engagement in AI science more accessible.
This goes beyond the issue of data gaps and data or human created algorithmic bias. Unlike the many other efforts looking at ethical AI, we are also focusing on the treatment of small minorities and outliers in any statistical determination of probability.
I suggested the reverse Wordle as a tool to get people thinking. It would also be good to get out the word and demonstrate how AI bias auditing tools do not detect bias against small minorities and outliers, such as people with disabilities, and that certification using these tools will do harm. I would like to explore on-device intelligence to maintain privacy and collect individualized data. It would be great if we looked at more promising small, thick data experiments and AI systems intended to work with more diverse data sets.
I look forward to discussing these with you soon,
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Wecount