After feeling stuck with the concept of subconscious biases and not knowing how to take it further, while I was feeling defeated, I needed to remind myself that this is a common experience. 
I was struggling to take my theoretical concept and turn it into a literal idea and concept. I recognised that addressing subconscious biases can be overwhelming and challenging, especially if they are deeply ingrained in someone's beliefs and behaviours. But despite this, the importance of

acknowledging personal bias is crucial to creating a more equitable society. And this is what I want to achieve throughout my project, it might not always be comfortable and I think for us to learn, we might need it to be uncomfortable as that will evoke us to really think about our actions as addressing our subconscious biases requires ongoing effort and self-reflection. 
After seeking guidance from a tutor, we came up with the idea of creating a sort of Tinder-like bias test that the audience would participate in, using an AI programme to generate faces. For some of these faces, I wanted to create in-depth descriptions of specific stereotypes while for others I would just create generic descriptions. 
I like the idea of using AI to generate the faces, as I think it  will be interesting to consider the possible bias already integrated into the AI generating programmes. 
At this time I wanted to create something interactive as I believe this is one of the best ways to educate ourselves about our own biases we hold. But I was concerned about there not necessarily being an outcome and ending with this idea if the audience partakes and 'swipes right' or 'swipes left' to vote on someone's face as whether they are a 'good' or 'bad' person, there still wouldn't be a conclusion that would make you question your biases. 
I went ahead with generating faces and stereotypes in the hope that an idea for possible data collecting would come to me as I was working.