In Part 1 of this series I laid out my starting thoughts on trusting the output from LLMs. In this article, I present a diagram I believe visualizes the trust “ecosystem”
Before sharing the diagram, I want to share two major themes within the diagram. . The first theme is contained within the definition of “trust” from Dictionary.com
The reliance on the integrity, strength, ability, surety, etc., of a person or thing; confidence.
Source: https://www.dictionary.com/browse/trust
Notice how it mentions a person OR thing ? It is the dynamics of human trust vs bot trust we are going to explore.
The second theme is associated with the time it takes to establish trust. As a part of writing this blog entry I found the following article entitled “10 Ways To Build Trust in a Relationship”. I am not suggesting that all 10 apply to Chatbots (or at least not yet) but I did find alignment with the following dimension:
Build Trust Gradually
Take Small Steps
Don’t expect too much too soon
Source: https://positivepsychology.com/build-trust/
So with those two themes explained - I can now move on to unveiling the diagram!
Introducing the “Keep Pace Visualizing ChatBot Trust” Diagram
One of my favorite phrases is “It’s exactly the same - but completely different!” As you will see, that phrase was definitely an influence in creating Version 1.0 of my “Visualizing Chatbot Trust” Diagram.
NOTE: You will be seeing this diagram several times throughout this “Trust Series” - as I will use it as my visual guide.
How the diagram is constructed:
My recommendation for reading this diagram is from the bottom up:
Your Trust Profile - Represents any individual - that in any situation is influenced by their knowledge and previous experiences with a given persona.
Trust Type - The idea is that when two humans interact - the dynamics are that of 2-way trust. With today’s LLM models -the bot is more at your command - thus the trust is one-way
Persona’s - There are a range of personas an individual will communicate with. For now this is broken into to broad categories - Human Beings and ChatBots. Will will dive deeper into specific later.
Trust Attributes - These are the various elements that are unique to each persona you engage with. It is that persona’s background that will influence what you do/don’t trust going forward.
So that concludes Part 2 !!! Would love to hear your thoughts !!!!
Upcoming Events I will be presenting at:
Every Sunday Morning (9AM EST) - Keep Pace Weekly Webinar on ChatGPT
Streaming on
www.facebook.com/keeppace
www.twitter.com/keeppace
www.linkedin.com/in/keeppace
May 18, 2023 - Artificial Intelligence - Innovation Summits on OpenAI & ChatGPT
Register for FREE at: https://www.linkedin.com/events/artificialintelligence7029930418099404800/comments/