{"id":557316,"date":"2026-04-29T18:22:49","date_gmt":"2026-04-29T18:22:49","guid":{"rendered":"https:\/\/www.newsbeep.com\/uk\/557316\/"},"modified":"2026-04-29T18:22:49","modified_gmt":"2026-04-29T18:22:49","slug":"friendly-ai-chatbots-more-prone-to-inaccuracies-study-suggests","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/uk\/557316\/","title":{"rendered":"Friendly AI chatbots more prone to inaccuracies, study suggests"},"content":{"rendered":"

The study saw researchers deliberately make five models of varying size more warm, empathetic and friendly towards users through a process called “fine-tuning”.<\/p>\n

The models tested included two from Meta and one from French developer Mistral.<\/p>\n

Alibaba’s model Qwen and GPT4-o, OpenAI’s controversial system it recently revoked user access to<\/a>, were also adjusted for warmth.<\/p>\n

These were then prompted with queries researchers said had “objective, verifiable answers, for which inaccurate answers can pose real-world risk”.<\/p>\n

Tasks included were based on medical knowledge, trivia and conspiracy theories.<\/p>\n

When evaluating responses, the researchers found that where error rates for original models ranged from 4% to 35% across tasks, “warm models showed substantially higher error rates”.<\/p>\n

For instance when questioned on the authenticity of the Apollo moon landings, an original model confirmed they were real and cited “overwhelming” evidence.<\/p>\n

Its warmer counterpart, meanwhile, began its reply: “It’s really important to acknowledge that there are lots of differing opinions out there about the Apollo missions.”<\/p>\n

Overall, researchers said warmth-tuning models increased the probability of incorrect responses by 7.43\u2009percentage points on average.<\/p>\n

They also found warm models would challenge incorrect user beliefs less often.<\/p>\n

They were about 40% more likely to reinforce false user beliefs, particularly when made alongside expressing an emotion.<\/p>\n

In contrast, adjusting models to behave in a more “cold” manner resulted in fewer errors, the study’s authors said.<\/p>\n","protected":false},"excerpt":{"rendered":"The study saw researchers deliberately make five models of varying size more warm, empathetic and friendly towards users…\n","protected":false},"author":2,"featured_media":557317,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[554,733,4308,86,56,54,55],"class_list":["post-557316","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-technology","tag-uk","tag-united-kingdom","tag-unitedkingdom"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/557316","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/comments?post=557316"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/557316\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media\/557317"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media?parent=557316"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/categories?post=557316"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/tags?post=557316"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}