I appreciate your focus on ethics. I’ve been exploring the Data Nutrition Project—tools that help assess bias in datasets. Worth a look!

My view on this has shifted a bit. I used to feel we should mostly let AI develop organically and deal with problems as they cropped up.

I still think we need lots of room for experimentation, but I now recognise that, because AI works at such scale and can quietly reinforce bias, we do need some shared safeguards from the outset – things like basic safety, non-discrimination and transparency about when AI is being used. Beyond that, I’d rather keep the space open for people and organisations to express their own values and be creative.

I’m still learning, but that’s where my thinking has got to so far.