A new job: Research Fellow in User-Centred Innovation for Accountable AI

Today marks my first day as a post-doctoral Research Fellow at the University of Aberdeen where I am working in the School of Natural and Computing Science as part of an interdisciplinary, EPSRC-funded project called RAInS (Realising Accountable Intelligent Systems).

Working with the RAInS project team and external stakeholders, my primary responsibilities for the next two years will be to investigate issues of trust and accountability in intelligent systems (AI). This will be done using qualitative research methods to answer questions surrounding the accountability of systems related to specific use cases.

For the non-AI folks, you can think of this (very, very simply) as looking at who/what is “responsible” when a self-driving car goes rouge or when AI-based hiring decisions are racist/sexiest.

The research was originally planned as a series of hands-on co-design workshops, but given the current state of the world (COVID19), we might need to re-think how best to proceed. I have been thinking about potential (digital) methods over the last couple of weeks and I am sure that we will have a few conversations about how to proceed in the next couple of weeks.

For now, I am just getting myself up to speed. I have several documents and literature to review to provide me a good overview of the project from background literature and proposals to research summaries and workshop artefacts from other research partners. I am hoping that by the time I get through those files we will have a better idea of how to proceed.

It is a bit of a strange start, given that we are on lockdown orders as part of the UK’s COVID19 response. This means that my work will be done remotely for the foreseeable future. I just hope that it won’t be long until I can move into my flat in Aberdeen and start meeting my new colleagues face-to-face. In the meantime, thank goodness for technology!

Leave a Reply

Your email address will not be published. Required fields are marked *