Sign up to see the future, today<\/p>\n
Can\u2019t-miss innovations from the bleeding edge of science and tech<\/p>\n
It doesn\u2019t take much to impress an AI chatbot.<\/p>\n
Tools like OpenAI\u2019s ChatGPT have long garnered a reputation<\/a> for being ludicrously sycophantic. Despite AI companies publicly promising<\/a> to address the problem, researchers recently found<\/a> that the bots still have a strong tendency to flatter and affirm in response to virtually any kind of prompt.<\/p>\n In the latest preposterous example<\/a> of this impulse, philosophy YouTuber and writer Jonas \u010ceika \u201csent ChatGPT an audio file of a series of FART sound effects and asked what it thinks of \u2018my music.’\u201d<\/p>\n It didn\u2019t take long for the glazing chatbot to congratulate him on his musical achievement, in what it called a \u201cstraight\u201d and \u201chonest reaction.\u201d<\/p>\n \u201cFirst impression: It has a cool lo-fi, late-night, slightly eerie vibe,\u201d it wrote. \u201cIt feels more like an atmosphere piece than a traditional song \u2014 which actually works in its favor. It reminds me of something that would play over a quiet city montage or end credits.\u201d<\/p>\n The bizarre feedback highlights how AI model sycophancy remains a considerable problem \u2014 and how \u201cChatGPT\u2019s musical analysis stinks!\u201d as the hosts of the \u201cPod Save America\u201d podcast joked during a recent episode<\/a>.<\/p>\n It certainly wouldn\u2019t be the first time an AI chatbot was caught giving outrageously misleading advice. Earlier this month, for instance, a TikTok user who goes by the moniker Husk asked ChatGPT in a viral video<\/a> to start a timer while he went to run a mile. When he told it to stop the timer just seconds later, the AI confidently told him he took over ten minutes to cover the distance.<\/p>\n While offering a musical analysis for fart sound effects may sound like a harmless gag, the tech\u2019s tendency to hallucinate and mislead could have far more serious consequences as well. For one, researchers warn that sycophantic conversations with AI could lull users into a potentially dangerous sense of intimacy and trust, facilitating anything from \u201cAI psychosis<\/a>\u201d and self harm<\/a> to acts of violence<\/a> in extreme cases.<\/p>\n More on hallucinations: Frontier AI Models Are Doing Something Absolutely Bizarre When Asked to Diagnose Medical X-Rays<\/a><\/p>\n \t