The fact that your bot generated valid explanations for the image is irrelevant to the thesis presented in the thread here, which is that chatbots are often engineered to use personal sounding language in order to activate the same behaviour in humans where we start thinking about who wrote the message. This is something we're evolved to do and something that can be played to get the user of a chatbot to develop a "relationship" with an unfeeling statistical prediction machine.
Prof. Emily M. Bender(she/her) (npub15xv…0jms)
