From Rob Fafoglia:

Absolutely. I talked about this in the session around what AI needs to be when it comes to collecting the training data. But just to expand on that a little bit, the way we keep it ethical is that we have to be ethical about how we source the training data and there are lots of different ways that we can do that. I talked about informed consent, so we have to make sure that people understand what we're using this data for, how it's going to be used and why we need it, and give people the option to opt out. If they don't want to take part then they don't have to. We want to make sure that we're protecting peoples privacy, that they understand that their data is going to be held in the strictest confidence and that we're not going to give up their personal and confidential information. We have to make sure that we're representing the population fairly. We're using people from all different walks of life, all sorts of different demographics, so we have a very diverse group of people that we're sourcing the data from. We want to make sure that we're not exploiting people, we're not collecting the data in any sort of manipulative way. We're respecting peoples dignity, autonomy and finally when we do the validation and checks, that that's ethical as well. We use the same ethical, responsible and diverse processes to review the data and make sure that it meets all that criteria.
So it really is in the way we collect it and the way we verify it that we continue to ensure that it's ethical.