_ghostKit is a platform that allows people to customise the personalities of their conversational interface systems.
Across the world there are millions of conversational devices in use, all deployed with standardized, built-in personalities – giving users no agency over how to relate to their systems. This concept is a layer over those systems, and works by linking device accounts, which would then sync personality preferences. Within each personality, the emotional states can be fine-tuned: humor, joy, sadness, sarcasm, anger, etc.
This leads to an elevated perception of an existing system – whether it is a nervous driver being encouraged by the voice on his GPS, or someone being informed about the weather with a poem. Every mundane interaction can hold a touch of magic, and a machine intelligence can start to feel more like a presence in your midst. A presence that can adjust to the situation, and make its users feel comfortable.
Initially aimed at providing a sense of companionship for socially isolated people, this project evolved into an exploration of the emotional relationship between humans and technology. With research informed by a psychologist, an AI personality designer, an anthropologist and AI rights activist, and a cognitive scientist who specialises in machine consciousness – this concept has been moulded from many perspectives.
The initial prototypes include a poetry bot trained on life’s work of seven poets, and a Volatile Personality Machine – which changes the tone of text, by reverse-engineering sentiment analysis APIs to do sentiment synthesis; and a paper prototype of the _ghostKit platform.
The result was a way to humanise the technology around us, because we are tempted by machines that offer companionship.