The need for greater diversity in AI algorithms

Yes, of course because we get better versions of AI, AI that doesn’t make mistakes, AI that represents all society, rather than AI that’s representing one chunk of the society. AI’s our next big step that can have a massive impact on how we do things and if we now decide that we’re going to have it, it’s going to be inclusive, universal, it’s going to cover people in rich countries, poor countries, people with different views on society. It will allow us to create an AI that’s representative to everyone and not simply representative to a few people who are developing it now. So, it is a massive step if we can involve everyone in this AI activity, rather than just certain groups.

So, of course, it benefits everybody if we have more people from third world poorer countries in that process, if we have more females in that process, if we have more minorities in that process and we would create a better AI that is going to work better for the people that it’s designed for. If we’re creating a system an intelligent agent that’s meant to help people in poor countries because they haven’t got enough doctors or whatever, it seems ridiculous that it’s been created by rich, white people when it’s working for non-rich middle class, middle aged possibly people.

So, we need input from everyone so we’re not answering, interpreting questions that need to be dealt with in our basis but for the people it’s going to be used and I know ever since the 1960’s when we started creating computer systems we wanted the user involvement but I think more than anything the development for AI is going to require user involvement at a level that we have not seen before.