Carefully reading the column by
Kevin Roose about his conversation with "Sydney" and her possible AI hallucinations; all this for a book chapter I am working on this summer.

https://www.nytimes.com/2023/02/16/technology/bing-chatbot-microsoft-chatgpt.html

I was thinking that story could lead us to think that these AI hallucinations perhaps have little to do with whether they are artificial constructions, imitations or actually symptoms of consciousness. That is, to relate these hallucinations with the normal state of consciousness in biological forms in which we all habitually experience similar symptoms. What differences would there be if self-critical awareness code is added to these Language models? A kind of virtual "superego". What would be the difference between that paradigm and ours? I know that I am not a specialist in the technical part of this artifacts, but I am trying to analyze it from a literary critical perspective. How would this interaction be analyzed if it were a science fiction text? Would the human character end up realizing that he/she is actually also a language model? #ArtificialIntelligence #AIHallucinations #scifi

@esfoobar Based on the Hungarian chatGPT I am an astronomer (no), having an Astronomy News YT channel (no), and wrote 4 books on astronomy (none). The English chatGPT says, I must be insignificant as he knows nothing about me. The English one is much closer to reality. At least chatGPT does not start hallucinating if it does not have at least a minimal information on the topic available. #chatGPT #aihallucinations #hungary

I continue to caution people using #OpenAI to generate content or answers on topics that the individual has surface level knowledge of. Out of curiosity I asked it to generate an “about” section of a stormcutter dragon from the #HowToTrainYourDragon series. There’s enough content on the internet for it to do so. It just generated an entirely fictional description of a #dragon with feathers and iridescent scales. And it imagined the dragon to appear in the first movie as well. #AIHallucinations

According to Perplexity AI, a new search interface that uses OpenAI #GPT 3.5 and Microsoft Bing, I won a #NobelPrize in Economics in 1983 (when I was born, link: https://www.perplexity.ai/?uuid=87677f47-c0e6-4f86-a71c-ac4288f2422c), a Nobel Prize in Physics in 2004 (link: https://www.perplexity.ai/?uuid=d9986bbc-af46-483e-84af-cc2afb86583d) and I have been President of Italy since 1962 Link: https://www.perplexity.ai/?uuid=e683d942-8bb7-44c1-874f-44a338e40d6b!

#languagemodels #StochasticParrots #AI #CognitiveIlusions #AIHallucinations #CognitiveScience #cognitiveAI