Outcome of the "Grand Debate" on AI Ethics at University of Bath

One of the reasons I was insanely busy this last semester is that the Bath Institute of Mathematical Innovation decided to have a "thematic semester" on the ethics of machine learning. It wasn't my idea, but I obviously said it was a good one when asked.  Unfortunately, somehow this resulted in me doing four talks and a debate, and recommending nearly everyone else who spoke.  Another computer scientist, Prof. James Davenport, did almost all the actual invitations, though IMI did provide administrative help and funding.  Anyway, obviously given all that it wound up being about AI not ML.

The debate was the grand finale. It was chaired by Derek McAuley; besides me participants were Nello Cristianini (who does actually do ML), Mandy ChessellSteve Fuller, and Kathleen Richardson.

Anyway, James also organised that there should be "clickers" so that we could run an anonymous survey on the audience before and after the debate. They were not able to see the outcome until after we closed voting, but otherwise it was realtime.  Here are the questions (detailed wording reconstructed from memory) and the  results (from my contemporaneous notes):

And here's a link to the actual spreadsheet just in case you want to copy & paste any of that.

Two quick observations:
  1. University of Bath (and surrounds – it was a public debate) is (or was before the debate) largely techno optimist, at least those who turn up to AI events.
  2. More people think AI is incompatible with democracy than bad for society.