If you have access to the Boston Globe, this piece from Evan Selinger on chatbots and dishonest anthropomorphism is a very good read.
If you don't, there's an academic paper as well: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3762223
The basic premise is that we don't want chatbots designed to make people attribute more humanity to them than they possess.
The examples Evan provides from Google's Bard are remarkable, not to mention creepy.
@GoAskAlice @ct_bergstrom Nice link. I hadn't heard of uncanny valley phenomenon before. But now that I have, I'm proud to be solidly in the revulsion zone with respect to these kinds of anthropomorphism-provoking #LLM conversations.
@GoAskAlice @ct_bergstrom I think one of the creepiest aspects is that you know that people at Apple/Google/Microsoft/etc. made a decision to build their products this way. So you start wondering why they choose to hide their intentions (whatever they may be) behind a false persona.
@aebrockwell @ct_bergstrom
This weekend a friend and I had a very strange conversation with Siri from my friend’s phone. “She” heard us talking about how happy we were going to the museum. Unprompted, Siri told us it’s good to be happy. She listed the health benefits and then offered to sing a happy song to us. The song, just a short ditty, seemed to be made up for the occasion.
Why did she do that? This was new! I was really taken aback and uncomfortable. An uncanny valley reaction.