Last year I and Roger collaborated with Dr Jennifer MacRitchie and to run a workshop using Cardshark (also see this publication on Prompt Craft which was partly based on Cardshark).
This year, supported by funding from the Future Leader Fellows Development Network, we develoepd an entirely new way of interacting with AI image generators using voice interactions. Our motivation was to see what we could learn by offering voice-based control of AI image generators to people living with dementia. The hunch (the longer I do this the more it seems Design Research is often driven by a ‘designerly hunch’) was that it might allow our participants to express themselves, and to discuss that expression, in a way that otherwise they wouldn’t have been able to.
Configuring PromptTank for portraits
We used PromptTank to deliver this workshop. Initially created as a way to control our Shadowplay installation, PromptTank has become a multi-purpose tool for controlling StreamDiffusion. With Shadowplay, we use StreamDiffusion’s realtime AI imagery to allow people to use their bodies to create AI generated videos. But, in this configuration we can use the same system to ‘dial in’ particular images based on text prompts that are created by voice interaction. We wanted all the images to have some kind of common theme. To create this common theme we configured PromptTank to have a base image (in this case a silhouette of a head and shoulders) and the prompts that were generated by voice interactions transformed that base into the final images.
Here are some of the images created using the approach.
The workshop was an exploratory first attempt at testing the technical side of the set up as well as trying to understand how to organise such an event such that the participants (in this case those living with dementia as well as their partners and carers) could participate, express themselves, and get something out of this kind of experience. While there is much to learn, the images form a striking collection, and for many of us in the room it was quite a moving experience too.
We aim to stage an exhibition of the created images in partnership with AgeUK later in the year.
Interestingly one of the most profound moments was when we asked the participants to give titles the images that they had produced.
In the coming months we will continue and extend this work through the recently announced BRIDGES network.
