[{"data":1,"prerenderedAt":338},["ShallowReactive",2],{"locale-alternates:\u002Fai-editorial-disclosure":3,"post-\u002Fai-editorial-disclosure":8},{"path":4,"alternates":5},"\u002Fai-editorial-disclosure",{"en":4,"tr":6,"de":7},"\u002Ftr\u002Fyapay-zeka-destegi-notu-okura-ne-anlatiyor","\u002Fde\u002Fki-kennzeichnung-was-leser-ueber-den-beitrag-des-autors-erfaehrt",{"page":9,"translations":190,"nav":196,"related":322,"random":325},{"id":10,"title":11,"body":12,"categories":162,"category":165,"changeHistory":165,"date":166,"description":167,"disclosures":168,"draft":171,"extension":172,"firstLiveAt":165,"image":173,"imageAlt":174,"kind":175,"lang":176,"meta":177,"navigation":178,"omitGermanLocalizationDisclosure":171,"path":4,"publishedAt":165,"readingTime":179,"rights":165,"seo":180,"seoTitle":181,"slug":182,"sources":165,"stem":182,"tags":183,"translationKey":182,"type":188,"updated":165,"__hash__":189},"posts\u002Fai-editorial-disclosure.md","What an AI Disclosure Tells Readers About the Author’s Contribution",{"type":13,"value":14,"toc":153},"minimark",[15,19,22,25,28,33,36,39,48,51,55,69,72,75,78,81,85,88,91,94,97,105,109,112,115,118,124,129,132,140,144,147,150],[16,17,18],"p",{},"Imagine seeing a note beside an article that says, “Prepared with AI assistance.” Did AI polish the author’s own prose? Did it turn the author’s notes into a first draft? Or did the model largely decide the subject, examples, and conclusion as well?",[16,20,21],{},"You have read the disclosure, but you still do not know how the work was done.",[16,23,24],{},"I use AI when preparing my articles, and I disclose that use. I have no measurement showing how these disclosures affect my readers’ trust.",[16,26,27],{},"I can still examine what the disclosure communicates. When I say, “I used AI,” I am providing information about a tool. A reader may be more interested in what that tool did while the thinking and the text took shape.",[29,30,32],"h2",{"id":31},"one-label-can-describe-very-different-writing-processes","One label can describe very different writing processes",[16,34,35],{},"One author develops the main idea, selects the examples, and writes the article. They then ask AI to simplify long sentences. Another author puts their views, experience, and sources into working notes and asks a model to produce the first draft. They then work through that draft.",[16,37,38],{},"The human contributes in both cases, but at a different stage. In the first, AI changes the expression of an existing text. In the second, it also produces the initial expression. “AI assistance” does not explain that distinction on its own.",[16,40,41,42,47],{},"My ",[43,44,46],"a",{"href":45},"\u002Fai-use-policy","AI Use Policy"," distinguishes language editing from more substantial drafting assistance. A general policy, however, cannot always explain what happened in each article. AI support might be limited to research in one piece. In another, the model might produce a substantial part of the first draft.",[16,49,50],{},"Nor can the difference be reduced to who wrote how many words. A person may write most of an article yet take its central claim from a model without challenging it. Conversely, detailed reasoning drawn from an author’s own experience may become a more readable article with a model’s help. Word count alone does not reveal where the thinking came from.",[29,52,54],{"id":53},"when-the-disclosure-is-vague-readers-fill-in-the-gap","When the disclosure is vague, readers fill in the gap",[16,56,57,58,68],{},"A 2024 PNAS Nexus ",[43,59,67],{"href":60,"className":61,"rel":63,"target":66},"https:\u002F\u002Fdoi.org\u002F10.1093\u002Fpnasnexus\u002Fpgae403",[62],"dofollow",[64,65],"nofollow","noopener","_blank","study by Sacha Altay and Fabrizio Gilardi on AI labels attached to news headlines"," examined this ambiguity in a controlled experiment. Online participants in the United States and the United Kingdom rated news headlines presented as social media posts.",[16,70,71],{},"In the second experiment, the label “Text generated by artificial intelligence” remained constant. The researchers explained that label through different divisions of work. One group was told that a journalist had selected the topic and written the article while AI edited its language and style. Another was told that the journalist had supplied the topic and sources while AI wrote the first draft. In a third explanation, AI selected the topic and produced the entire article. A further group saw the label without an explanation.",[16,73,74],{},"The researchers considered ratings of headline accuracy together with people’s willingness to share them. Explanations describing language editing and first-draft production from supplied sources produced more positive results than the unexplained label. Scores in those two groups were not significantly higher than those in the unlabeled control group.",[16,76,77],{},"What changed was the division of work described to participants. The explanations were also given in advance, checked for comprehension, and repeated. The experiment therefore does not show that adding a somewhat longer note to a web page will always produce better results. It also measured immediate ratings of headlines, not the news items’ actual accuracy or whether participants shared them in practice.",[16,79,80],{},"My narrower conclusion is this: before predicting how readers will respond, we need to be clear about which writing process we are describing to them.",[29,82,84],{"id":83},"readers-may-expect-more-from-an-author-than-accurate-information","Readers may expect more from an author than accurate information",[16,86,87],{},"Consider an executive reading another executive’s account of a difficult decision. The reader may be looking for accurate information. They may also want to understand which option that person rejected and why, what they misjudged, and how they faced the outcome.",[16,89,90],{},"Firsthand experience matters to that expectation. “I made this decision” promises something different from “This is a decision someone could make in such a situation.” The second can be useful analysis. But it can replace the first only if the experience being described actually happened.",[16,92,93],{},"Treating every reader who cares about AI use as inherently biased misses this distinction. A reader can question the accuracy of the information while also wanting to understand the author’s contribution. Whether an idea was developed by the author, relayed from someone else, or built from a model’s suggestion may matter to why the reader is spending time on the article.",[16,95,96],{},"Research has not established this as true of every reader. It is an expectation I consider reasonable in the relationship between author and reader. Telling readers, “If the information is accurate, why should its production matter?” would mean deciding on their behalf what they are allowed to value.",[16,98,99,100,104],{},"In my earlier article on ",[43,101,103],{"href":102},"\u002Fwhy-we-trust-ai-judgments","why we trust AI judgments",", I distinguished accepting a recommendation from making a sound decision. Here, the question is what readers expect from the person whose name appears on the article. A text can contain accurate information and still fail to meet that expectation.",[29,106,108],{"id":107},"a-useful-disclosure-makes-the-division-of-work-visible","A useful disclosure makes the division of work visible",[16,110,111],{},"Taking responsibility for an article does not describe all the work that went into preparing it. Approving a text and writing its first draft are different contributions. So are finding a source and deciding what conclusion that source supports.",[16,113,114],{},"An AI-use note may therefore be more useful when it describes the model’s main contribution rather than naming the model or counting how many times it was used. What material did the author provide? Which part did AI prepare? What did the author subsequently examine and change?",[16,116,117],{},"Consider two hypothetical disclosures. They do not document how every article on this site was produced:",[119,120,121],"blockquote",{},[16,122,123],{},"The author’s text was edited with AI assistance to simplify the language and correct errors. The author reviewed the final text.",[119,125,126],{},[16,127,128],{},"The first draft was prepared with AI from the author’s views, experience notes, and sources. The author checked whether the sources supported the relevant claims, revised the text, and approved the final version.",[16,130,131],{},"If the checks in the second note did not actually happen, those sentences cannot be used. A more detailed disclosure can simply become a more detailed claim. Its value comes from matching the work that was done.",[16,133,134,135,139],{},"In my ",[43,136,138],{"href":137},"\u002Fopen-publishing-ai-monthly-seo-reporting-experiment","Open Publishing and monthly SEO reporting experiment",", I separately describe the process behind AI-assisted assessments. A report and a personal account may not need the same disclosure. This distinction is useful because it shows how the work was actually done, not because it makes the process look more human or more automated than it was.",[29,141,143],{"id":142},"disclosure-should-inform-readers-not-persuade-them","Disclosure should inform readers, not persuade them",[16,145,146],{},"I would not judge an AI disclosure by whether everyone who sees it likes the article to the same degree. A clearer note may reassure some readers. Others may decide that the contribution it describes is not what they were looking for. I have not measured which response my own readers have.",[16,148,149],{},"The disclosure should give them enough information to make that decision. If I used AI for language editing, it should say so. If a model prepared the first draft, it should say that instead. If human review took place, the note should explain what was reviewed and should not imply that a check occurred when it did not.",[16,151,152],{},"The name on an article tells readers whom they are engaging with. The disclosure can help them understand how that person contributed to the text. That is the value of moving from “I used AI” to “This is what AI did in this article, and this is what I contributed.”",{"title":154,"searchDepth":155,"depth":155,"links":156},"",2,[157,158,159,160,161],{"id":31,"depth":155,"text":32},{"id":53,"depth":155,"text":54},{"id":83,"depth":155,"text":84},{"id":107,"depth":155,"text":108},{"id":142,"depth":155,"text":143},[163,164],"ai","business",null,"2026-09-30","Saying that AI was used in an article does not explain who shaped its ideas. A useful disclosure shows what the author and the model each contributed.",{"aiUse":169,"aiNote":170},"ai-assisted","Evren Bal determined the subject, personal context, and intended argument. AI assisted with source review, research synthesis, the first Turkish draft, and adaptation of this English version from the approved Turkish article.",false,"md","\u002Fimages\u002Fhero\u002Fai-editorial-disclosure.avif","An author lifts a disclosure sheet to reveal sources, AI assistance, and final human review.","Essay","en",{},true,6,{"title":11,"description":167},"AI Content Disclosure: What Did the Author Contribute?","ai-editorial-disclosure",[184,185,186,187],"ai-assisted-writing","authorship","editorial-transparency","ai-content-disclosure","post","IQJYfgu7jkSXlsBptBBcJIWIkxH1Tppj5Z45jMzxkbg",{"en":191,"tr":192,"de":194},{"path":4,"title":11},{"path":6,"title":193},"“Yapay zekâ desteğiyle hazırlandı” notu okura ne anlatıyor?",{"path":7,"title":195},"Was eine KI-Kennzeichnung über den Beitrag des Autors aussagt",{"prev":197,"next":165,"others":200,"lucky":319,"readingTime":179},{"path":198,"title":199},"\u002Fwhen-does-ai-progress-improve-everyday-life","When Does AI Progress Become Progress for People?",[201,204,207,210,213,216,219,222,225,228,231,234,237,240,243,246,249,252,255,258,261,264,267,270,273,276,279,282,285,288,291,294,295,298,301,304,307,310,313,316],{"path":202,"title":203},"\u002Fwhy-the-same-ai-model-produces-different-results","Why the Same AI Model Produces Different Results Across Applications",{"path":205,"title":206},"\u002Fmanaging-technology-and-transforming-the-business-are-not-the-same","Managing Technology and Transforming the Business Are Not the Same Thing",{"path":208,"title":209},"\u002Fif-ai-handles-the-execution-who-sets-the-strategy","If AI Handles the Execution, Who Sets the Strategy?",{"path":211,"title":212},"\u002Fhow-much-authority-should-ai-have","How Much Authority Should You Give an AI System?",{"path":214,"title":215},"\u002Fthe-job-ai-wont-take-and-the-five-it-prevents","AI Is Reducing Hiring Without Layoffs",{"path":217,"title":218},"\u002Fwhere-should-people-be-in-the-loop-while-ai-does-the-work","Where Should People Be in the Loop While AI Does the Work?",{"path":220,"title":221},"\u002Fraising-children-in-the-age-of-artificial-intelligence","Raising Children in the Age of Artificial Intelligence",{"path":223,"title":224},"\u002Fredar-ai-powered-summaries-for-kap-disclosures-and-open-sources","Redar: AI-Powered Summaries for KAP Disclosures and Open Sources",{"path":226,"title":227},"\u002Fwhat-does-the-eu-ai-act-actually-regulate","What Is the EU AI Act, and What Does It Regulate?",{"path":229,"title":230},"\u002Fis-your-ai-system-delivering-business-results","Is Your AI System Actually Delivering Business Results?",{"path":232,"title":233},"\u002Fhow-llms-identify-experts","What Makes an LLM Recommend Someone as an Expert?",{"path":235,"title":236},"\u002Fbeyond-the-bot-lessons-from-building-a-chat-system-for-global-patients","What We Learned Building a Healthcare Chatbot for International Patients",{"path":238,"title":239},"\u002Fbank-account-api-integration","Integrating One Bank Is Easy. Keeping Forty Running Is Not.",{"path":241,"title":242},"\u002Fan-seo-experiment-in-a-low-competition-serp-with-google-maps-and-openai","Building camiler.org: A Programmatic SEO Experiment with Google Maps and OpenAI",{"path":244,"title":245},"\u002Fai-turned-a-refactor-i-wouldnt-do-into-a-one-hour-job","With AI, a Refactor I Wouldn't Bother With Took One Hour",{"path":247,"title":248},"\u002Fai-knowledge-management-from-documents-to-better-decisions","Why Searching Company Knowledge with AI Is Not Enough",{"path":250,"title":251},"\u002Fai-smaller-teams-business-capacity","AI Makes Smaller Teams Possible. What Does the Business Gain — and Lose?",{"path":253,"title":254},"\u002Fwhen-does-enterprise-ai-need-rag","When Does Enterprise AI Actually Need RAG?",{"path":256,"title":257},"\u002Fai-made-code-cheap-verification-is-still-expensive","AI Made Code Cheap. Verification Is Still Expensive.",{"path":259,"title":260},"\u002Fhow-to-route-requests-across-multiple-ai-models","How Should You Route Requests Across Multiple AI Models?",{"path":262,"title":263},"\u002Fai-assisted-rest-api-development","Preserving API Quality in AI-Assisted Development",{"path":265,"title":266},"\u002Fthe-threshold-collapsed-to-zero","The Threshold Collapsed: What ProductLog Taught Me About Building in Public",{"path":268,"title":269},"\u002Fturkey-national-ai-platform-evren","National AI Infrastructure Is More Than a GPU Count",{"path":271,"title":272},"\u002Fhow-i-built-a-modern-infrastructure-using-open-source-tools-and-the-power-of-cloudflare","Why I Run camiler.org on a Single VPS",{"path":274,"title":275},"\u002Fai-tools-are-easier-to-build-transformation-still-takes-work","Building an AI Tool Is Easy. Transforming a Company Is Not.",{"path":277,"title":278},"\u002Fai-assisted-migrations-deferred-work","AI Can Make Deferred Work Worth Doing",{"path":280,"title":281},"\u002Fwhy-im-building-rankextension-making-google-search-console-actually-make-sense","What Happened to RankExtension?",{"path":283,"title":284},"\u002Fhow-to-do-content-pruning-a-real-world-case-study","How to Do Content Pruning: A Real-World Case Study",{"path":286,"title":287},"\u002Fai-transformation-redesign-work-not-cut-roles","AI Transformation Starts with Redesigning Work",{"path":289,"title":290},"\u002Fdont-write-to-be-recommended-by-ai","Getting Your Brand Recommended by AI Is the Wrong Goal. What Is the Right One?",{"path":292,"title":293},"\u002Fwhen-is-one-ai-model-enough","When Is One AI Model Enough?",{"path":198,"title":199},{"path":296,"title":297},"\u002Fwhy-sales-and-other-departments-keep-clashing","Why Sales and Other Departments Keep Clashing",{"path":299,"title":300},"\u002Fis-buying-backlinks-worth-it","Would We Still Buy This Backlink If It Sent No Ranking Signal?",{"path":302,"title":303},"\u002Fhealthcare-ai-chatbot-triage-medical-device","AI in Healthcare: The Difference Between Chatbots, Triage, and Medical Devices",{"path":305,"title":306},"\u002Fstart-with-the-business-problem-not-the-ai-model","Start With the Business Problem, Not the AI Model",{"path":308,"title":309},"\u002Fdo-you-know-how-dependent-your-company-is-on-ai","Do You Know How Dependent Your Company Is on AI?",{"path":311,"title":312},"\u002Fprotecting-organizational-memory-during-ai-transformation","Protect Organizational Memory During an AI Transformation",{"path":314,"title":315},"\u002Fdoes-enterprise-ai-really-need-fine-tuning","Does Enterprise AI Really Need Fine-Tuning?",{"path":317,"title":318},"\u002Fwriting-with-ai-means-thinking-with-your-archive","Writing with AI Means Thinking About Your Archive, Too",{"path":320,"title":321},"\u002Fai-visibility-is-also-a-state-concern","AI Visibility Is Also on States’ Agendas",[323],{"path":317,"title":318,"date":324},"2026-09-06",[326,330,334],{"path":327,"title":328,"date":329},"\u002Fwordpress-to-nuxt-ai-powered-content-pipeline","From WordPress to Nuxt: Building an AI-Powered Content Pipeline","2026-06-22",{"path":331,"title":332,"date":333},"\u002Fllms-txt-was-never-the-point","llms.txt Was Never the Point","2026-08-23",{"path":335,"title":336,"date":337},"\u002Fhow-much-does-the-eu-ai-act-affect-companies-outside-the-eu","How Much Does the EU AI Act Affect Companies Outside the EU?","2026-09-07",1790742069697]