The cutting edge of AI, dubbed \u201cPhD level\u201d intelligence by its creators, thinks there are three Rs in the words \u201cNorthern Territory\u201d.<\/p>\n
That is what users discovered after trying out the latest advancement of OpenAI, GPT-5.<\/p>\n
During a GPT-5 launch event on Thursday, the OpenAI CEO, Sam Altman, described the latest version of ChatGPT<\/a> as like having \u201caccess to a PhD-level expert in your pocket\u201d, comparing the previous version to a college student, and the one before that a high school student.<\/p>\n When users on social media attempted to challenge GPT-5, however, they found the model making basic errors in its responses. One Bluesky user<\/a> found the chatbot repeatedly claimed there were three Bs in \u201cblueberry\u201d.<\/p>\n \u201cYep \u2013 blueberry is one of those words where the middle almost trips you up, like it\u2019s saying \u2018b-b-better pay attention,\u2019\u201d the chatbot said in the posted chat.<\/p>\n Sign up: AU Breaking News email<\/a><\/p>\n \u201cThat little bb moment is satisfying, though \u2013 it makes the word feel extra bouncy.\u201d<\/p>\n Another user<\/a> found the chatbot unable to correctly identify the US states containing the letter R. And when asked to produce a map, it misspelled states including \u201cKrizona\u201d and \u201cVermoni\u201d. ChatGPT also double listed California and invented the states \u201cNew Jefst\u201d and \u201cMitroinia\u201d.<\/p>\n When Guardian Australia asked<\/a> the latest model of ChatGPT to identify the number of Rs in Australia\u2019s states and territories, it could identify those which did. But the AI also believed Northern Territory had just three Rs not five. When asked to produce it on a map, spelled the territory as \u201cNorthan Territor\u201d<\/p>\n OpenAI was contacted for comment. The company said on launching the new product that there would be fewer errors and AI hallucinations under the latest model.<\/p>\n The source of the issues may be the way GPT-5 works as a combination of its AI models.<\/p>\n GPT-5 has a \u201creal-time router\u201d that quickly decides which of its model to use depending on the conversation type and intent. OpenAI has said that if you ask ChatGPT<\/a> to \u201cthink hard about this\u201d it uses the latest reasoning model.<\/p>\n \u201cThe router is continuously trained on real signals, including when users switch models, preference rates for responses, and measured correctness, improving over time,\u201d OpenAI\u2019s announcement stated.<\/p>\n skip past newsletter promotion<\/a><\/p>\n Sign up to Breaking News Australia<\/p>\n Get the most important news as it breaks<\/p>\n Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy<\/a>. We use Google reCaptcha to protect our website and the Google Privacy Policy<\/a> and Terms of Service<\/a> apply.<\/p>\n after newsletter promotion<\/p>\n The chief executive of media and AI startup Every, Dan Shipper, found that the model would sometimes hallucinate on questions posed to it that should have been routed to the reasoning model.<\/p>\n