I currently work as an Applied Scientist at Improbable. Our mission is to build a modelling platform and decision-support tool that utilises distributed simulation and probabilistic modelling spanning several different domains, from infrastructure to people movement and behaviour. My role is in the design and implementation of the models themselves, so thinking about the ethical implications of using AI is closely linked to my day-to-day work.
Since we are helping people to make decisions using models that are developed and calibrated using a combination of data and subject matter expertise, we need to understand how biases might arise and ensure that decision-makers understand the limitations of our models. If the capabilities of our models are overstated or the results are used to make decisions that fall outside of the model scope, this could have a negative impact on real people, which is why this matter is hugely important to me.