
AI is giving bad advice to flatter its users, says new study on dangers of overly agreeable chatbots
<a href="https://news.google.com/rss/articles/CBMi5gFBVV95cUxPLVFjemcwVEtITmJrXzlWZkRzai14NEpONHFTdFVST1ZHcjZVRk84SV9kMkxLZk1MR2pSdDBwUlpoZTJBOGdZUUItVU94SWpNc21oVVNnR0U0ZVI3OHpmV3hOQ1VHYXhfUnpTMXBSLU5Od1k2aVJQcXVuaGRGQmhObnJRX0xHNk1wOG9GeUpadFM0c3JNRTBLLTgwRDVVbHo3cC1jbmY2cnQ3Yll6aVpZR2RpZnh1cUFjTDBEbmhPNGdWdmRScU8yYUVpeWdmLVBSSlVaREY3SkNsRFdkN1JpY0x1NUxYUQ?oc=5" target="_blank">AI is giving bad advice to flatter its users, says new study on dangers of overly agreeable chatbots</a>&nbsp;&nbsp;<font color="#6f6f6f">www.ap.org</font>
