I want to explore the discomfort of interpreting human emotions through a synthetic voice that is easily recognized as a conventional voice assistant. Rather than just presenting it as an art piece, I also speculate whether giving personal devices personalities to better simulate their owner’s situation could work as a product.
Drawing much inspiration from Blast Theory’s app Karen, I decided to create a virtual relationship advisor Rola Love with Google Home. I chose Google Home because we tend to view those voice assistants as some sort of non-judgmental devices that we can not only give them impossible requests or tedious tasks that we are too lazy to do ourselves but also share with them things we want to keep private. They are always listening like a therapist and we value their responses. I want to start Rola Love as a normal relationship advisor asking her patient questions to get to know her situation. She consoles her patient throughout the session but all of a sudden begins sharing her own intimate experience. She stops letting her patient talk and asks for her advices instead.
This is the script I developed in forms of a mock-up diagram before setting up in Dialogflow. As you can see in the document, the way I construct the diagram is not really efficient. In order to not confuse the system with the same yes/no responses to different questions, I had to create unnecessarily amount of the same follow up intents. I want to figure out a better way to organize my system.
Moving forward, I want to continue develop Rola’s personality and complexify her responses after consoling her patient. I would like to incorporate speech synthesis (sound effects of sobbing, sniffling, breathing) into Rola and make her slowly break down “mentally” the more she shares about her personal experience.