Jungian personality in a chatbot
Summary
We explore how to add Jungian personality to a chatbot as a process.
`Salve', an existing serious chatbot game,
was used as a starting point and closely inspected.
We designed personality as a preference for an algorithm rather than value
driven decision making, akin to the ideas of Campos.
In other words, a personality prefers a way of doing things rather than
the content they prefer.
With this in mind a strategy was devised for adding personality to the existing
chatbot in the `Salve' serious game,
while leaving as much of the original architecture in tact.
This caused us to replace the existing AIML based scheme with a novel
approach.
In this approach, we opened up space to have varied responses to a similar
utterance depending on the personality.
The Drools rule engine is now the center of deliberation in the chatbot.
The new modelling method was also less verbose and more precise.
We conclude by demonstrating that this scheme works as expected.