A few quick thoughts post-conf:
In nearly every session the phrase " Your model is only as good as your data" was uttered.
The only real work surrounding ML/AI is being done in the area of engineering - “inside the box”, everything “outside the box” (pre/post modelling) - e.g. design, planning, assurance - is barely acknowledged in either business or academia.
A general air of optimism pervaded the sessions “this WILL improve life” although most of the speakers were engineers who have nothing to do with the outcomes their work will bring.
Dr Lorien Pratt was the best speaker. Too much info to summarise but note-to-self to read all of her references starting with https://www.responsibleaidi.org/
In a nutshell: