IA and data science
I work with data scientists to add a user-centred and human lens to algorithms and mathematics.
Many machine learning and artificial intelligence projects aim to use data to improve decision-making and recommendations. Data-driven approaches can introduce bias caused by historical data or by using classifications that contain bias, and inadvertently exclude people (particularly groups that have historically been marginalised).
I work with data scientists and data specialists to make sure we are working with data and classifications in a way that reduces bias and provides the best outcome for people. That can involve:
- understanding biases in classifications and taxonomies that have been used to code historical data
- developing taxonomies that contain less bias
- developing criteria to understand what a ‘good’ outcome looks like
- cleaning up historical data
- incrementally improving models
I’m not a data scientist, but I understand mathematical concepts well (I dropped out of math after meeting unreal numbers in first year uni, and studied statistics instead) and understand modelling (that’s what an economics degree helps with). I apply my expertise in cognition, classification and language to mathematical and algorithmic concepts.