Litha Chatbot Technology… the early days!
In July 2019, Litha Chatbot technology spoke to our lead Psychologist Ian for the first time. Initial lab tests were through a Slack Channel but this was the first time via a smart speaker. Ian interacted with the Litha Chatbot via Google Assistant.
We ran the first chat under the name of ‘Eliza‘. The refers to an early natural language processing computer program created from in 1966 at the MIT AI Laboratory.
You can hear ‘mountain to climb’ – this is so that we have a marker about which part of the sentence Litha picks up. You can hear how Litha and Google Assistant talk about a mountain to climb 11 times.
To Listen & Understand
As with almost every single human conversation, the more abstract and complex the topic is, the more we use metaphors and imagery to explain a concept.
The aim for Litha is to listen and understand.
What you can also hear is how Litha is using our innovative platform to help you to understand where you are within a specific situation.
Whereas the majority of AI uses statistical probability (thinking, “in all likelihood, based on what I’m told…”), Litha is building technology focused on embracing ambiguity, meaning and context.
Humans think and reason in images, not statistics
The same applies with regards to the communications engine. It’s not your standard chat bot that has structured question & answer mechanisms (decision trees) and, again, looking at the statistical probability of what you’re saying (many bots break your sentences into ‘tokens’ identifying verb, advertising, subject, noun, etc. before deciding what you’re likely to be saying). Litha asks questions to gather the meaning – and, in doing so, satisfies a key aspect of psychotherapy which is to seek understanding and not give answers.
The transcript referred to at the end helps to form your own ‘self-authored narrative’ – an ongoing journal of your discussions and thoughts. Communicate however you like and Litha sends you a transcript to confirm what was discussed.
The evolution of Litha from here is, on the surface, simple: she will become more natural in her conversation and this will build quicker rapport with the user.