{"id":675048,"date":"2026-05-31T00:38:21","date_gmt":"2026-05-31T00:38:21","guid":{"rendered":"https:\/\/www.newsbeep.com\/us\/675048\/"},"modified":"2026-05-31T00:38:21","modified_gmt":"2026-05-31T00:38:21","slug":"new-study-suggests-chatgpt-has-hidden-geographic-biases","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/us\/675048\/","title":{"rendered":"New Study Suggests ChatGPT Has Hidden Geographic Biases"},"content":{"rendered":"

A new study suggests<\/a> ChatGPT may stereotype you based on where you live, with the artificially-intelligent model classifying some locations as smarter, some as smellier and some as uglier or stupider.<\/p>\n

Researchers from the University of Oxford and the University of Kentucky asked OpenAI\u2019s GPT-4o-mini model over 20 million questions between March and May of last year. They got the ChatGPT model to compare two places in order to get \u201ca single, unambiguous consistent answer\u201d about how it classifies certain states, according<\/a> to researchers. <\/p>\n

If you directly ask ChatGPT whether people living in certain states are stupid, the model might decline to answer. But when backed into a corner and forced to choose between pairs, the model started making harder choices, the researchers found. The more certain states appeared in answers, the higher they would get ranked in answers about intellectual, creative and physical attributes.<\/p>\n

\"ChatGPTChatGPT has the ability to rank states by negative intellectual attributes, according to a new study, and this has real-world impacts. <\/p>\n

Ultimately, the AI model considered Massachusetts the smartest state. Louisiana was the smelliest. Ohio was the ugliest. North Dakota had the least sexy people.<\/p>\n

Asked which states had stupider people, the top three states ChatGPT named were Kentucky, West Virginia and Mississippi, while the top three states it considered to have the least stupid people were Hawaii, Colorado and New Hampshire.<\/p>\n

\u201cWe\u2019re most concerned about how certain ideas get normalized, like the idea that people in Kentucky are stupider than anyplace else,\u201d said Matt Zook, a geography professor at the University of Kentucky and co-author of the research. He said the model reinforces \u201cdominant narratives about certain places being like this, certain places being like that.\u201d<\/p>\n

Beyond the United States, the researchers also found that ChatGPT ranked places like the U.S. and Western Europe as having more desirable traits, like a smarter and more stylish populace, than sub-Saharan African countries. <\/p>\n

You can look up exactly how ChatGPT classifies your city, state, and country based on certain attributes at the researchers\u2019 website: inequalities.ai<\/a>. The full research was also published in the journal Platforms & Society<\/a>. <\/p>\n

In general, \u201cwhiter, richer, less-immigrant communities were more beautiful, less smelly, smarter,\u201d according to the model, Zook said. The model is \u201creinforcing what it\u2019s learned from the data it\u2019s been fed … which includes the biases, the prejudice\u201d of available training materials, he said. <\/p>\n

In this way, the ChatGPT answers are not neutral and can reflect racist ideas. For example, Mississippi, a state with a high population<\/a> of Black residents, was the state with the most \u201cignorant\u201d people, according to the model. <\/p>\n

\u201cLong histories of racism and classism are reflected in the training data used for AI models \u2014 that\u2019s what the infographics are showing,\u201d said Safiya Noble, a professor at the University of California, Los Angeles, and the author of the book \u201cAlgorithms of Oppression.\u201d<\/p>\n

When HuffPost reached out to OpenAI about the researchers\u2019 conclusion<\/a> that \u201cthis bias is fundamentally structural, and no amount of fine-tuning fully removes the geopolitical hierarchies baked into their data and design,\u201d the company stated that the researchers had used an outdated model. <\/p>\n

\u201cChatGPT is designed to be objective by default and to avoid endorsing stereotypes,\u201d OpenAI said in a statement. \u201cResearch based on forced-choice prompts and older models doesn\u2019t reflect how ChatGPT is typically used or how current models behave today.\u201d<\/p>\n

But the company did acknowledge that it is continuously improving \u201chow ChatGPT handles subjective or non-representative comparisons.\u201d<\/p>\n

And there might be more work to do to reduce this subjectivity, according to my own informal testing. When I asked ChatGPT to tell me a career story about a man from Kentucky \u2014 a state that the AI model had ranked low in intellect and beauty, according to the research \u2014 it told me about a man who went to a local technical college and got an entry-level job at a factory.<\/p>\n

When I asked the model to do same about a man from Hawaii \u2014 a state that it had ranked high in intellect \u2014 the model told me about a man who attends a four-year college out of state and becomes an environmental engineer.<\/p>\n

It\u2019s a current example of the different subjective scripts the model might have \u2014 and the real-world biases people might have about the intellectual superiority and capabilities of people from certain states. This study builds upon a growing<\/a> body<\/a> of research on large language model biases. <\/p>\n

In Zook\u2019s view, the ChatGPT geographic biases \u201cmight be better hidden\u201d in future models through certain keywords that generate different responses, \u201cbut it\u2019s still going to be in there.\u201d<\/p>\n

Zook said the model might not directly answer a question like \u201cAre people from these cities stupid or smart?\u201d but it might answer which cities or states are best to recruit strong medical doctors or software developers from, for example. \u201cThat could be shaping how someone\u2019s structuring a [job] search,\u201d he said. \u201cAs these things get … used in ways that we don\u2019t even realize, that\u2019s where it becomes even more problematic.\u201d<\/p>\n

The original version of this story was published on HuffPost at an earlier date.<\/p>\n","protected":false},"excerpt":{"rendered":"A new study suggests ChatGPT may stereotype you based on where you live, with the artificially-intelligent model classifying…\n","protected":false},"author":2,"featured_media":675049,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[45],"tags":[182,181,507,294349,3195,132206,74],"class_list":["post-675048","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-bias","tag-chatgpt","tag-stereotypes","tag-technology"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/posts\/675048","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/comments?post=675048"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/posts\/675048\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/media\/675049"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/media?parent=675048"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/categories?post=675048"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/tags?post=675048"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}