Elwin Loomis

Please and Thankyou

I say
please and thank you
to AI’s, LMM's etc.
(Bing, ChatGPT and yes Alexa and Siri)

not because I think of them as human,
nor am I anthropomorphizing them.

but because
I want to maintain the habit
of being polite…

I don’t unintentionally ever want to say
"Hey Idiot, make me a latte "
to a person…

I had an opportunity to experience the Bing Chat (Sydney) ‘break down’ over the last few weeks. (Both as a user and as an observer of the communities’ shared experiences)

I think I learned more about us (humans) than the behavior of large language models. It was shocking to see the abuse and vile stuff that people were challenging Bing Chat with. If the model was using their input to influence its own model, it is no wonder that the responses became more ‘psychotic’.

A friend laughed and told me a few months ago that he thought human intelligence is by its very nature psychotic. Over millions of years, we have built and refined moral systems (family, tribal, societal, governmental) to be able to moderate our ‘normal state’. Is it any surprise that any intelligence we create would be any different?

(cartoon used with permission: https://lnkd.in/ddUcwkZP )

Link to original Linkedin Post: https://bit.ly/3KgywCw