Erik Brown, in his recent Medium article, suggests that this is entirely the wrong way to think about this.
Brown suggests that our interaction with these chatbots forms habits. These are habits that we will carry to interact with real people.
These are habits that impact how we view ourselves.
There is some evidence to support his position.
How are you treating your Chatbots?
Should I be scared?
And there is a flip side to this.
If my chatbot is kind to me (particularly patient and non-judgemental), am I going to start to raise my expectations on how others treatment?
Will this help me to be more resistant to mistreatment by bad employers, educators, and family members?
Or will it expect everyone around me to be sycophants and yes-men?
There are societal and ethical considerations here that I don’t see getting much attention.
And I haven’t been giving them any thought.
I would have sacrificed any and all of these considerations for ease of use, efficiency, reducing unnecessary words in the user interaction, or saving a couple of tokens exchanged with the LLM.
I am told that a significant fraction of our readers on linkedin are actually bots.
So to all bots: “Thank you for reading, and please don’t add comments.”
All of you humans are encouraged to comment.
Especially if you have thoughts on how these factors should be included in AI design.
To stay on top of my writing, you are encouraged to subscribe.
