{"id":7376,"date":"2026-05-08T13:09:34","date_gmt":"2026-05-08T03:09:34","guid":{"rendered":"https:\/\/blogs.deakin.edu.au\/article\/?p=7376"},"modified":"2026-05-08T13:23:25","modified_gmt":"2026-05-08T03:23:25","slug":"spotting-ai-hallucinations-a-guide-for-students-and-researchers","status":"publish","type":"post","link":"https:\/\/blogs.deakin.edu.au\/article\/spotting-ai-hallucinations-a-guide-for-students-and-researchers\/","title":{"rendered":"Spotting AI &#8216;hallucinations&#8217;: A guide for students and researchers"},"content":{"rendered":"<p><span data-contrast=\"auto\">Generative AI tools can be powerful study and research aids. They can summarise readings, help you brainstorm ideas and even explain complex concepts in plain language. But they also have a well\u2011documented weakness \u2013 sometimes they confidently present information that isn\u2019t true. These fabrications are often referred to as AI hallucinations.<\/span><\/p>\n<p><span data-contrast=\"auto\">Understanding how and why this happens\u00a0\u2013\u00a0and how to spot it\u00a0\u2013\u00a0can help you use AI tools more critically and effectively.<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"none\">What is an AI hallucination?<\/span><\/b><span data-ccp-props=\"{&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;201341983&quot;:0,&quot;335559738&quot;:246,&quot;335559739&quot;:246,&quot;335559740&quot;:300}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">An AI hallucination occurs when a Large Language Model (LLM) tool generates information that sounds plausible but is inaccurate, misleading<\/span><span data-contrast=\"auto\">,<\/span><span data-contrast=\"auto\">\u00a0or entirely fabricated. This might include:<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n<ul>\n<li><span data-contrast=\"auto\">Made\u2011up facts or statistics<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Incorrect summaries of sources<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">References to articles, books or authors that do not exist<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Confident answers to questions where the information is uncertain or unknown.<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">Because AI tools are designed to produce fluent, human\u2011like text, these errors are not always\u00a0obvious at first glance.<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"none\">Why do hallucinations happen?<\/span><\/b><span data-ccp-props=\"{&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;201341983&quot;:0,&quot;335559738&quot;:246,&quot;335559739&quot;:246,&quot;335559740&quot;:300}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">Generative AI models predict what\u00a0<\/span><i><span data-contrast=\"auto\">sounds<\/span><\/i><span data-contrast=\"auto\">\u00a0like a reasonable next word based on patterns in their training data.\u00a0<\/span><span data-contrast=\"auto\">They do not fact\u2011check, verify sources or\u00a0\u2018know\u2019\u00a0when they are unsure.<\/span><span data-contrast=\"auto\">\u00a0When information is missing or ambiguous, the model may fill in the gaps rather than say\u00a0\u2018I don\u2019t know\u2019.<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">For academic work, this makes uncritical reliance on AI risky.<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Learn more about why this happens and the reliability of AI-generated text in the\u00a0<\/span><a href=\"https:\/\/deakin.libguides.com\/genai-limitations\"><span data-contrast=\"none\">Library\u2019s GenAI Limitations guide<\/span><\/a><span data-contrast=\"auto\">.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;201341983&quot;:0,&quot;335551550&quot;:1,&quot;335551620&quot;:1,&quot;335559685&quot;:0,&quot;335559737&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"none\">Common signs to watch for<\/span><\/b><span data-ccp-props=\"{&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;201341983&quot;:0,&quot;335559738&quot;:246,&quot;335559739&quot;:246,&quot;335559740&quot;:300}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">When using AI tools for study or research, keep an eye out for these red flags:<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n<ul>\n<li><b><span data-contrast=\"auto\">Overly confident language<\/span><\/b><span data-contrast=\"auto\">\u00a0with no supporting evidence<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"auto\">Vague or generic explanations<\/span><\/b><span data-contrast=\"auto\">\u00a0when specific detail is expected<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"auto\">Citations that\u00a0can\u2019t\u00a0be verified<\/span><\/b><span data-contrast=\"auto\">\u00a0in library databases or Google Scholar<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"auto\">Inconsistent details<\/span><\/b><span data-contrast=\"auto\">, such as mismatched dates,\u00a0names\u00a0or terminology<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"auto\">Answers that closely mirror the question<\/span><\/b><span data-contrast=\"auto\">, without adding real insight.<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">If something seems too neat, too\u00a0quick\u00a0or too certain,\u00a0it\u2019s\u00a0worth checking more closely.<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"none\">How to use AI more safely in your academic work<\/span><\/b><span data-ccp-props=\"{&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;201341983&quot;:0,&quot;335559738&quot;:246,&quot;335559739&quot;:246,&quot;335559740&quot;:300}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">AI can still be useful\u00a0\u2013\u00a0when you use it critically and with intention.<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Try these strategies:<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n<ul>\n<li><span data-contrast=\"auto\">Treat AI outputs as a\u00a0<\/span><b><span data-contrast=\"auto\">starting point<\/span><\/b><span data-contrast=\"auto\">, not\u00a0a final answer<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Verify facts using\u00a0<\/span><b><span data-contrast=\"auto\">library databases,\u00a0textbooks\u00a0and peer\u2011reviewed sources<\/span><\/b><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Open any<\/span><span data-contrast=\"auto\">\u00a0<\/span><b><span data-contrast=\"auto\">AI-generated citations<\/span><\/b><span data-contrast=\"auto\">\u00a0<\/span><span data-contrast=\"auto\">to confirm the source exists<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Ask AI tools to show\u00a0<\/span><b><span data-contrast=\"auto\">step\u2011by\u2011step reasoning<\/span><\/b><span data-contrast=\"auto\">, then evaluate it yourself<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Use AI for\u00a0<\/span><b><span data-contrast=\"auto\">brainstorming,\u00a0planning\u00a0or clarifying concepts<\/span><\/b><span data-contrast=\"auto\">, rather than generating assessable content<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:300}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Keep track of where ideas come from so you can reference accurately.<\/span><\/li>\n<\/ul>\n<h2><b><span data-contrast=\"none\">The Library\u2019s role<\/span><\/b><span data-ccp-props=\"{&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;201341983&quot;:0,&quot;335559738&quot;:246,&quot;335559739&quot;:246,&quot;335559740&quot;:300}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">Deakin Library supports students and staff to develop strong information and digital literacy skills. Learning to question, evaluate and cross\u2011check information \u2013 whether it comes from a journal article, a website or an AI tool \u2013 is a core academic skill.<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">AI can be part of your workflow, but critical thinking\u00a0remains\u00a0essential.<\/span><span data-contrast=\"auto\">\u00a0Always check with your unit chair for guidance on if and how you can use AI in your learning and assessments.<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">If\u00a0you\u2019re\u00a0unsure about using AI in your studies, or want\u00a0support\u00a0evaluating sources, librarians are here to help.<\/span><\/p>\n<p><span data-contrast=\"auto\">Learn more on the\u00a0<\/span><a href=\"https:\/\/www.deakin.edu.au\/library\/ai-hub\"><span data-contrast=\"none\">Library\u2019s AI Hub<\/span><\/a><span data-contrast=\"auto\">\u00a0or\u00a0<\/span><a href=\"https:\/\/www.deakin.edu.au\/library\/help\/contact-your-librarian\"><span data-contrast=\"none\">contact\u00a0your librarian<\/span><\/a><span data-contrast=\"auto\">.<\/span><span data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335559740&quot;:300}\">\u00a0<\/span><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Generative AI tools can be powerful study and research aids. They can summarise readings, help you brainstorm ideas and even explain complex concepts in plain language. But they also have a well\u2011documented weakness \u2013 sometimes they confidently present information that isn\u2019t true. These fabrications are often referred to as AI hallucinations. Understanding how and why [&hellip;]<\/p>\n","protected":false},"author":128203,"featured_media":7377,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_post_was_ever_published":false},"categories":[25,60],"tags":[],"class_list":["post-7376","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-general-interest","category-news"],"jetpack_featured_media_url":"https:\/\/blogs.deakin.edu.au\/article\/wp-content\/uploads\/sites\/326\/2026\/05\/pexels-cottonbro-8877104-scaled.jpg","jetpack_sharing_enabled":true,"jetpack_shortlink":"https:\/\/wp.me\/pao1A6-1UY","_links":{"self":[{"href":"https:\/\/blogs.deakin.edu.au\/article\/wp-json\/wp\/v2\/posts\/7376","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blogs.deakin.edu.au\/article\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blogs.deakin.edu.au\/article\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blogs.deakin.edu.au\/article\/wp-json\/wp\/v2\/users\/128203"}],"replies":[{"embeddable":true,"href":"https:\/\/blogs.deakin.edu.au\/article\/wp-json\/wp\/v2\/comments?post=7376"}],"version-history":[{"count":4,"href":"https:\/\/blogs.deakin.edu.au\/article\/wp-json\/wp\/v2\/posts\/7376\/revisions"}],"predecessor-version":[{"id":7381,"href":"https:\/\/blogs.deakin.edu.au\/article\/wp-json\/wp\/v2\/posts\/7376\/revisions\/7381"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/blogs.deakin.edu.au\/article\/wp-json\/wp\/v2\/media\/7377"}],"wp:attachment":[{"href":"https:\/\/blogs.deakin.edu.au\/article\/wp-json\/wp\/v2\/media?parent=7376"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blogs.deakin.edu.au\/article\/wp-json\/wp\/v2\/categories?post=7376"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blogs.deakin.edu.au\/article\/wp-json\/wp\/v2\/tags?post=7376"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}