{"course":{"slug":"ai-for-students","title":"AI for Students","url":"https://addaly.com/learn/ai-for-students","level":"beginner","track":"work"},"lesson":{"slug":"research-and-real-sources","title":"Research, And Sources That Exist","url":"https://addaly.com/learn/ai-for-students/research-and-real-sources","position":38,"of":73,"minutes":8,"answer":"Never cite a source you have not opened. Asking the model if it is real does not count.","keyIdea":"Never cite a source you have not opened. Asking the model if it is real does not count."},"body":{"format":"markdown","text":"## The failure that keeps happening\n\nIn March 2023, a court filing written by New York lawyer Steven Schwartz and signed by his colleague Peter LoDuca cited six court decisions. Names, docket numbers, judges, quoted passages. All six were invented by ChatGPT.\n\nThe part worth memorising is what he did when opposing counsel could not find them. He went back to the chatbot and asked whether the cases were real. It said yes. He asked for confirmation. It produced full text. Excerpts of it were filed with the court too. In June 2023 the judge fined the two lawyers and their firm $5,000, jointly.\n\nStudents do the smaller version of this every term, and it is far easier to catch than AI-written prose, because a marker who wants to check a reference checks it. A fabricated bibliography is one of the few reliable signals a tired academic has.\n\n## Why citations in particular\n\nA language model, absent a search tool, is producing statistically plausible text. A citation is an unusually learnable format: author surname, year, a title made of field-appropriate nouns, a journal that publishes in that field, a volume number, a page range, a DOI-shaped string. Every component is drawn from a pattern the model has seen thousands of times.\n\nThe result is a reference that is right in every respect except existing.\n\nThis is not the model lying. There is no database being consulted and no lookup being fumbled. Unless the tool is actually searching, and you should always know whether yours is, nothing in the process ever checks reality.\n\n## The rule, and how to satisfy it\n\n**Never cite a source you have not opened.** Not summarised, not confirmed. Opened.\n\nHow to check, in order, ninety seconds each:\n\n1. **Search the exact title** in Google Scholar, your library catalogue, or your discipline's index: PubMed, JSTOR, SSRN, arXiv. If the exact title returns nothing anywhere, stop. It does not exist.\n2. **Resolve the DOI.** Paste it after `https://doi.org/`. A real DOI lands on the paper. An invented one returns an error. This takes five seconds and catches most fabrications.\n3. **Check author, year, journal and volume together.** Fabrications frequently attach real, well-known researchers to a paper they never wrote. Finding that Bjork is a real memory researcher tells you nothing about whether this particular Bjork paper exists.\n4. **Open it and find your sentence.** This catches the more common failure, which is not invention at all: a real paper cited for a claim it does not make, or makes about a different population, or explicitly argues against.\n\nThat last one is worth dwelling on. Genuine papers get miscited more often than fake ones get invented, and miscitation is the error that survives all the way into your finished work.\n\n## Search-enabled tools change the risk, not the rule\n\nMore tools now retrieve and link: ChatGPT with browsing, Perplexity, Gemini, Claude with search, Copilot. A link you can click is different from a fabricated reference, and this is real progress.\n\nBut the summary can still drift from the source. The tool retrieves a page, compresses it, and the compression may sharpen a hedged finding into a firm one, or lose the sample size, or attribute a result to the wrong study among several on the page. Click through. Read the paragraph. It is quick, and it is the difference between citing and gesturing.\n\nAlso notice what the retrieved sources actually are. A confident answer built on three blog posts and a press release is not the same as one built on the paper.\n\n## Where AI helps with research\n\nUse it for the map, not the territory.\n\n```text\nWhat are the main competing positions on whether minimum\nwage increases reduce employment? Who argues each side,\nand what is the strongest objection to each?\n```\n\n```text\nI'm researching informal waste collection in Indian cities.\nWhat search terms would a subject librarian use that I\nprobably wouldn't think of?\n```\n\n```text\nHere's the methods section of a paper. Explain the\nregression discontinuity design in plain terms so I can\nfollow the results.\n```\n\nAll three are things it is good at, and none produce a citation. You get orientation, vocabulary, and comprehension; then you go and read.\n\nOne more, which will save you an afternoon: after you have your source list, ask it what you appear to have missed, or which viewpoint is absent from your reading. A bibliography that only contains people who agree with you is a common, avoidable mark deduction."},"check":{"question":"Your draft cites a paper with a plausible title, a real journal, volume and page numbers, and a DOI. What actually establishes that it exists?","options":[{"text":"Asking the model to confirm the reference is genuine and to supply the abstract.","correct":false,"tempting":"It will confirm, at length and convincingly: this is precisely the step that got a New York lawyer sanctioned in 2023."},{"text":"The presence of a well-formed DOI, since DOIs are issued by registries and cannot simply be made up.","correct":false,"tempting":"The format looks institutional and authoritative, but a model can emit a DOI-shaped string that resolves to nothing at all."},{"text":"Confirming the listed authors are real researchers who publish in that field.","correct":false,"tempting":"Real researchers are routinely attached to papers they never wrote, so verifying the people leaves the paper unverified."},{"text":"Resolving the DOI or locating the record in a library or index, then opening it and finding the sentence you say it supports.","correct":true}]},"prev":{"slug":"essays-across-subjects","title":"The essay subjects, and what a marker is reading for","url":"https://addaly.com/learn/ai-for-students/essays-across-subjects"},"next":{"slug":"how-a-citation-gets-invented","title":"How a citation gets invented","url":"https://addaly.com/learn/ai-for-students/how-a-citation-gets-invented"},"authorship":"Founder and series editor: Dr Mustafa Mun. Written by AI models to a written standard. Exam keys checked blind by a second model: 3,665 of 3,666 agreed. Academic review invited.","review":{"status":"Draft, open for academic review"},"cite":"Addaly (2026) 'Research, And Sources That Exist', in AI for Students. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/ai-for-students/research-and-real-sources (Accessed: 31 October 2026).","licence":{"name":"CC BY-NC-SA 4.0","url":"https://creativecommons.org/licenses/by-nc-sa/4.0/"},"reader":{"schema":1,"title":"Research, And Sources That Exist","courseTitle":"AI for Students","module":{"index":5,"slug":"research-and-evidence","title":"Research, sources and evidence"},"answer":null,"keyIdea":"Never cite a source you have not opened. Asking the model if it is real does not count.","hero":{"kind":"plate","image":null,"alt":null,"caption":null},"blocks":[{"t":"h2","id":"the-failure-that-keeps-happening","runs":[{"text":"The failure that keeps happening"}]},{"t":"p","runs":[{"text":"In March 2023, a court filing written by New York lawyer Steven Schwartz and signed by his colleague Peter LoDuca cited six court decisions. Names, docket numbers, judges, quoted passages. All six were invented by ChatGPT."}]},{"t":"p","runs":[{"text":"The part worth memorising is what he did when opposing counsel could not find them. He went back to the chatbot and asked whether the cases were real. It said yes. He asked for confirmation. It produced full text. Excerpts of it were filed with the court too. In June 2023 the judge fined the two lawyers and their firm $5,000, jointly."}]},{"t":"p","runs":[{"text":"Students do the smaller version of this every term, and it is far easier to catch than AI-written prose, because a marker who wants to check a reference checks it. A fabricated bibliography is one of the few reliable signals a tired academic has."}]},{"t":"h2","id":"why-citations-in-particular","runs":[{"text":"Why citations in particular"}]},{"t":"p","runs":[{"text":"A language model, absent a search tool, is producing statistically plausible text. A citation is an unusually learnable format: author surname, year, a title made of field-appropriate nouns, a journal that publishes in that field, a volume number, a page range, a DOI-shaped string. Every component is drawn from a pattern the model has seen thousands of times."}]},{"t":"p","runs":[{"text":"The result is a reference that is right in every respect except existing."}]},{"t":"p","runs":[{"text":"This is not the model lying. There is no database being consulted and no lookup being fumbled. Unless the tool is actually searching, and you should always know whether yours is, nothing in the process ever checks reality."}]},{"t":"h2","id":"the-rule-and-how-to-satisfy-it","runs":[{"text":"The rule, and how to satisfy it"}]},{"t":"p","runs":[{"text":"Never cite a source you have not opened.","b":true},{"text":" Not summarised, not confirmed. Opened."}]},{"t":"p","runs":[{"text":"How to check, in order, ninety seconds each:"}]},{"t":"ol","start":1,"items":[[{"text":"Search the exact title","b":true},{"text":" in Google Scholar, your library catalogue, or your discipline’s index: PubMed, JSTOR, SSRN, arXiv. If the exact title returns nothing anywhere, stop. It does not exist."}],[{"text":"Resolve the DOI.","b":true},{"text":" Paste it after "},{"text":"https://doi.org/","code":true},{"text":". A real DOI lands on the paper. An invented one returns an error. This takes five seconds and catches most fabrications."}],[{"text":"Check author, year, journal and volume together.","b":true},{"text":" Fabrications frequently attach real, well-known researchers to a paper they never wrote. Finding that Bjork is a real memory researcher tells you nothing about whether this particular Bjork paper exists."}],[{"text":"Open it and find your sentence.","b":true},{"text":" This catches the more common failure, which is not invention at all: a real paper cited for a claim it does not make, or makes about a different population, or explicitly argues against."}]]},{"t":"p","runs":[{"text":"That last one is worth dwelling on. Genuine papers get miscited more often than fake ones get invented, and miscitation is the error that survives all the way into your finished work."}]},{"t":"h2","id":"search-enabled-tools-change-the-risk-not-the-rule","runs":[{"text":"Search-enabled tools change the risk, not the rule"}]},{"t":"p","runs":[{"text":"More tools now retrieve and link: ChatGPT with browsing, Perplexity, Gemini, Claude with search, Copilot. A link you can click is different from a fabricated reference, and this is real progress."}]},{"t":"p","runs":[{"text":"But the summary can still drift from the source. The tool retrieves a page, compresses it, and the compression may sharpen a hedged finding into a firm one, or lose the sample size, or attribute a result to the wrong study among several on the page. Click through. Read the paragraph. It is quick, and it is the difference between citing and gesturing."}]},{"t":"p","runs":[{"text":"Also notice what the retrieved sources actually are. A confident answer built on three blog posts and a press release is not the same as one built on the paper."}]},{"t":"h2","id":"where-ai-helps-with-research","runs":[{"text":"Where AI helps with research"}]},{"t":"p","runs":[{"text":"Use it for the map, not the territory."}]},{"t":"code","lang":"text","label":"Text","text":"What are the main competing positions on whether minimum\nwage increases reduce employment? Who argues each side,\nand what is the strongest objection to each?"},{"t":"code","lang":"text","label":"Text","text":"I'm researching informal waste collection in Indian cities.\nWhat search terms would a subject librarian use that I\nprobably wouldn't think of?"},{"t":"code","lang":"text","label":"Text","text":"Here's the methods section of a paper. Explain the\nregression discontinuity design in plain terms so I can\nfollow the results."},{"t":"p","runs":[{"text":"All three are things it is good at, and none produce a citation. You get orientation, vocabulary, and comprehension; then you go and read."}]},{"t":"p","runs":[{"text":"One more, which will save you an afternoon: after you have your source list, ask it what you appear to have missed, or which viewpoint is absent from your reading. A bibliography that only contains people who agree with you is a common, avoidable mark deduction."}]}],"check":{"question":[{"text":"Your draft cites a paper with a plausible title, a real journal, volume and page numbers, and a DOI. What actually establishes that it exists?"}],"options":[{"text":[{"text":"Asking the model to confirm the reference is genuine and to supply the abstract."}],"correct":false,"tempting":[{"text":"It will confirm, at length and convincingly: this is precisely the step that got a New York lawyer sanctioned in 2023."}]},{"text":[{"text":"The presence of a well-formed DOI, since DOIs are issued by registries and cannot simply be made up."}],"correct":false,"tempting":[{"text":"The format looks institutional and authoritative, but a model can emit a DOI-shaped string that resolves to nothing at all."}]},{"text":[{"text":"Confirming the listed authors are real researchers who publish in that field."}],"correct":false,"tempting":[{"text":"Real researchers are routinely attached to papers they never wrote, so verifying the people leaves the paper unverified."}]},{"text":[{"text":"Resolving the DOI or locating the record in a library or index, then opening it and finding the sentence you say it supports."}],"correct":true,"tempting":null}],"order":[1,3,2,0],"why":null},"prev":{"slug":"essays-across-subjects","title":"The essay subjects, and what a marker is reading for","minutes":8},"next":{"slug":"how-a-citation-gets-invented","title":"How a citation gets invented","minutes":8}}}