{"id":40716,"date":"2026-09-16T06:46:40","date_gmt":"2026-09-16T06:46:40","guid":{"rendered":"https:\/\/smartdev.com\/?p=40716"},"modified":"2026-09-16T06:46:40","modified_gmt":"2026-09-16T06:46:40","slug":"when-ai-gets-fraud-detection-wrong-who-is-accountable","status":"publish","type":"post","link":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/","title":{"rendered":"When AI Gets Fraud Detection Wrong: Who Is Accountable?"},"content":{"rendered":"<div id=\"fws_6aaab7d71a21d\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<p dir=\"ltr\"><strong>TL;DR<\/strong><\/p>\n<ul dir=\"ltr\">\n<li>AI generates risk signals, not decisions &#8211; organizations choose what those signals are allowed to trigger.<\/li>\n<li>&#8220;Wrong&#8221; isn&#8217;t just a bad prediction: false positives, false negatives, and turning a probability into an automatic action all cause harm differently.<\/li>\n<li>Responsibility gets spread across many teams, but accountability still has to sit with one named decision owner.<\/li>\n<li>A human reviewer in the workflow isn&#8217;t real oversight unless they have information, authority, and enough time to actually judge.<\/li>\n<li>The goal is AI-assisted judgment, not autonomous decisions: AI detects and assembles evidence, humans decide on consequential cases.<\/li>\n<\/ul>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-40717 size-full\" src=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_24_24-AM.png\" alt=\"\" width=\"1536\" height=\"1024\" srcset=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_24_24-AM.png 1536w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_24_24-AM-300x200.png 300w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_24_24-AM-1024x683.png 1024w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_24_24-AM-768x512.png 768w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_24_24-AM-18x12.png 18w\" sizes=\"auto, (max-width: 1536px) 100vw, 1536px\" \/><\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71a4cb\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<h3 aria-level=\"3\"><span class=\"ez-toc-section\" id=\"Introduction\"><\/span><b><span data-contrast=\"none\">Introduction\u00a0<\/span><\/b><span data-ccp-props=\"{&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;335559738&quot;:160,&quot;335559739&quot;:80}\">\u00a0<\/span><span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><span data-contrast=\"auto\">Across\u00a0<\/span><a href=\"https:\/\/smartdev.com\/fr\/ai-in-finance-top-use-cases-and-real-world-applications\/\"><span data-contrast=\"none\">AI use cases in finance<\/span><\/a><span data-contrast=\"auto\">, fraud detection is one of the areas where machine learning has moved fastest from pilot to production. AI now helps fraud teams handle a scale and speed of data that humans simply\u00a0can&#8217;t\u00a0match on their own.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">The numbers explain why. By 2025, Thomson Reuters Institute found that\u00a0<\/span><a href=\"https:\/\/thepaymentsassociation.org\/article\/ai-and-fraud-prevention-the-hidden-risks-of-false-positives-and-black-box-models\/\"><span data-contrast=\"none\">71% of financial services firms were already using AI for risk assessment and reporting<\/span><\/a><span data-contrast=\"auto\">, second only to document summarization. The pressure to adopt AI is real. Global fraud losses reached an estimated \u00a31.03 trillion in 2024. Close to half the world&#8217;s population reported encountering at least one\u00a0scam\u00a0per week during that same period.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Still, speed and scale\u00a0aren&#8217;t\u00a0the whole story. Consider what happens when an AI-driven decision causes\u00a0real harm: blocking a transaction, freezing an account, or holding back a seller&#8217;s payout. Or consider the opposite failure, where the system lets a genuine fraud slip through. In either case, the question\u00a0isn&#8217;t\u00a0just &#8220;was the model accurate?&#8221; The bigger question is:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p style=\"text-align: center;\"><i><span data-contrast=\"auto\">When AI influences a fraud decision, who owns the outcome?<\/span><\/i><span data-ccp-props=\"{&quot;335551550&quot;:2,&quot;335551620&quot;:2}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">The thesis of this piece is simple:\u00a0AI can automate detection, but accountability cannot be\u00a0automated away.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71a675\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<h3 aria-level=\"3\"><span class=\"ez-toc-section\" id=\"What_Does_It_Mean_for_AI_to_%E2%80%9CGet_Fraud_Detection_Wrong%E2%80%9D\"><\/span><b><span data-contrast=\"none\">What Does It Mean for AI to &#8220;Get Fraud Detection Wrong&#8221;?<\/span><\/b><span data-ccp-props=\"{&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;335559738&quot;:160,&quot;335559739&quot;:80}\">\u00a0<\/span><span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><span data-contrast=\"auto\">Before talking about accountability, it helps to define what &#8220;wrong&#8221; actually means.\u00a0As the next section shows,\u00a0the failure\u00a0rarely sits with the model alone.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">This pattern\u00a0isn&#8217;t\u00a0unique to fraud, either. It\u00a0shows up\u00a0wherever teams treat AI outputs as final answers rather than inputs to a decision.\u00a0We&#8217;ve\u00a0explored the same dynamic in the context of\u00a0<\/span><a href=\"https:\/\/smartdev.com\/fr\/when-ai-gets-compliance-wrong-the-hidden-risk-of-hallucination\/\"><span data-contrast=\"none\">compliance hallucination<\/span><\/a><span data-contrast=\"auto\">, where a confidently wrong AI output can pass for a verified fact.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71a7fd\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<h4 aria-level=\"4\"><b><span data-contrast=\"none\">&#8220;Wrong&#8221; is not just a wrong prediction<\/span><\/b><span data-ccp-props=\"{&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;335559738&quot;:80,&quot;335559739&quot;:40}\">\u00a0<\/span><\/h4>\n<p><span data-contrast=\"auto\">Three distinct types of failure are worth separating.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">False positive\u00a0&#8211;\u00a0legitimate behavior gets classified as fraud. The consequences here are not trivial. Systems block legitimate transactions.\u00a0Accounts get frozen. Customers get forced through unnecessary verification steps. Trust and conversion take a hit. This\u00a0isn&#8217;t\u00a0a theoretical risk, either. A 2025 global survey found\u00a0that roughly\u00a0<\/span><a href=\"https:\/\/antivirusinsider.com\/fraud-detection-and-prevention-statistics\/\"><span data-contrast=\"none\">6 in 10 ecommerce merchants reported false-positive rates between 2% and 10% on disputed orders<\/span><\/a><span data-contrast=\"auto\">. Global losses from false declines &#8211; legitimate orders wrongly rejected &#8211; reached an estimated<\/span><\/p>\n<p><span data-contrast=\"auto\">$201 billion\u00a0in 2025.\u00a0That&#8217;s\u00a0roughly 1.51%\u00a0of annual ecommerce revenue lost purely to over-cautious systems. Even more telling, 33% of consumers who experienced a false decline said they would not shop with that business again. In other words, a false positive\u00a0isn&#8217;t\u00a0just an operational cost.\u00a0It&#8217;s\u00a0a\u00a0trust\u00a0cost.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">False negative\u00a0&#8211;\u00a0fraudulent activity passes through undetected. Here, the consequences\u00a0show up as\u00a0direct financial losses, chargebacks, compliance exposure, and repeat fraud. In ecommerce specifically, friendly fraud and account takeover together account for\u00a0<\/span><a href=\"https:\/\/antivirusinsider.com\/fraud-detection-and-prevention-statistics\/\"><span data-contrast=\"none\">64% of all recorded fraud cases in 2026<\/span><\/a><span data-contrast=\"auto\">. This signals a shift: more sophisticated, harder-to-catch fraud types now dominate over the crude patterns that rule-based systems were originally built to catch.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">A wrong action taken from an uncertain prediction.\u00a0This is the failure mode most worth digging\u00a0into, because\u00a0teams overlook it most often. AI rarely\u00a0states\u00a0outright, &#8220;this transaction is fraud.&#8221;\u00a0Instead, it produces a probability, something like:\u00a0<\/span><i><span data-contrast=\"auto\">&#8220;There is an 82% probability this transaction is suspicious.&#8221;<\/span><\/i><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">The real problem lives in the next step. Crucially, humans decide this step, not the model:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p style=\"text-align: center;\"><em>82% risk score \u2192 automatically\u00a0block\u00a0the transaction\u00a0<\/em><\/p>\n<p><span data-contrast=\"auto\">That business rule turns a probabilistic signal into a definitive action. As a result, the system can &#8220;get it wrong&#8221; even when the model, technically, does exactly what its designers built it to do.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71a997\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<h4 aria-level=\"4\"><b><span data-contrast=\"none\">Why these failures happen<\/span><\/b><span data-ccp-props=\"{&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;335559738&quot;:80,&quot;335559739&quot;:40}\">\u00a0<\/span><\/h4>\n<p><span data-contrast=\"auto\">Rather than listing ten scattered causes, it helps to name three underlying ones.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ul>\n<li><span data-contrast=\"auto\">Reality keeps changing. Fraud tactics, customer behavior, and transaction patterns shift constantly. This causes model drift: a model trained on past data gradually loses accuracy as the world it models moves on.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">AI sees patterns, not full context. An anomaly might signal fraud. Or it might just reflect unusual but entirely legitimate behavior, like a customer using their card abroad right after booking a flight online.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Prediction turns into action too quickly. The biggest mistake often has nothing to do with the model. Instead, it lies in how an organization converts Signal \u2192 Decision \u2192 Action, frequently without a real checkpoint in between.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-40722 size-full\" src=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_34_53-AM.png\" alt=\"\" width=\"1536\" height=\"1024\" srcset=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_34_53-AM.png 1536w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_34_53-AM-300x200.png 300w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_34_53-AM-1024x683.png 1024w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_34_53-AM-768x512.png 768w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_34_53-AM-18x12.png 18w\" sizes=\"auto, (max-width: 1536px) 100vw, 1536px\" \/><\/span><\/p>\n<p><span data-contrast=\"auto\">Together, these three causes point to the core insight of this piece:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Fraud detection models generate risk signals.\u00a0Organizations decide what those signals are allowed to trigger.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><span data-contrast=\"auto\">This insight is also the bridge into accountability.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71ab25\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<h3 aria-level=\"3\"><span class=\"ez-toc-section\" id=\"The_Accountability_Gap_Everyone_Touches_the_System_But_Who_Owns_the_Decision\"><\/span><b><span data-contrast=\"none\">The Accountability Gap: Everyone Touches the System, But Who Owns the Decision?<\/span><\/b><span data-ccp-props=\"{&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;335559738&quot;:160,&quot;335559739&quot;:80}\">\u00a0<\/span><span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><span data-contrast=\"auto\">In practice, an AI fraud system involves a wide cast of stakeholders, and each one touches only a slice of the outcome. AI and Data Science teams build the model, tune it, and retrain it as new fraud patterns\u00a0emerge. Fraud Operations teams work the alerts day to day, deciding in the moment which cases deserve escalation. Product and Engineering teams ship the system, wire it into checkout flows or account-management dashboards, and decide how alerts surface to reviewers.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Risk and Compliance teams, meanwhile,\u00a0set\u00a0the governance framework: acceptable risk appetite, regulatory obligations, and escalation policy. Management approves strategy and signs off on budget and scope. Often, a third-party vendor supplies the underlying model, the training data, or both. That vendor&#8217;s own design choices &#8211; which the buying organization rarely audits line by line &#8211; end up shaping outcomes just as much as any internal decision does.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Herein lies the problem: many hands share responsibility, yet accountability blurs across all of them. Everyone plays a role, and everyone can point to someone else&#8217;s piece of the system when something goes wrong. The data scientist can say the model performed within expected tolerances. The analyst can say they followed the alert as presented. The product team can say they built what compliance specified. Compliance can say the vendor&#8217;s model made the call. Each statement can be true on its own, and the customer whose account got frozen still has no clear answer to &#8220;who decided this?&#8221;<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71acc8\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<p><span data-contrast=\"auto\">To untangle this, it helps to separate four distinct layers of responsibility, since collapsing them into one vague notion of &#8220;the AI team&#8221; is exactly what lets accountability slip through the cracks:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-40723 size-full\" src=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_42_05-AM.png\" alt=\"\" width=\"1536\" height=\"1024\" srcset=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_42_05-AM.png 1536w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_42_05-AM-300x200.png 300w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_42_05-AM-1024x683.png 1024w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_42_05-AM-768x512.png 768w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_42_05-AM-18x12.png 18w\" sizes=\"auto, (max-width: 1536px) 100vw, 1536px\" \/><\/p>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"7\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Technical responsibility\u00a0&#8211;\u00a0Who is accountable for model performance (accuracy, drift, bias)? This covers whether the model does what it was designed to do, at the level of precision and recall the organization agreed to accept.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"8\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Operational responsibility\u00a0&#8211;\u00a0Who actually uses the alerts and recommendations day to day?\u00a0This covers how consistently staff follow, override, or escalate what the model surfaces.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"9\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Governance responsibility\u00a0&#8211;\u00a0Who sets the thresholds, escalation rules, and acceptable risk levels?\u00a0This is where a probability score gets translated into a business action, and where most of the real risk actually lives.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"10\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Decision accountability\u00a0&#8211;\u00a0Who is responsible when a consequential action actually happens?\u00a0This is the layer that ultimately\u00a0answers to\u00a0the customer, the regulator, and the board.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71ae8e\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<p class=\"PDq2pG_selectionAnchorContainer\" data-start=\"81\" data-end=\"305\">Mapping a single incident against these four layers usually reveals the gap immediately. A model might perform exactly as validated (layer 1 is clean). An analyst might also follow the alert as designed (layer 2 is clean).<\/p>\n<p data-start=\"310\" data-end=\"582\">Yet, the threshold that turned an 82% risk score into an automatic account freeze might never have gone through a formal risk review (layer 3 fails). No single role might have signed off on \u201cwe will auto-freeze accounts at this confidence level\u201d (layer 4 never existed).<\/p>\n<p data-start=\"587\" data-end=\"725\">In cases like this, no individual technically did anything wrong. Yet the organization still produced a harmful and unaccountable outcome.<\/p>\n<p><span data-contrast=\"auto\">The point to land on:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p style=\"text-align: center;\"><i><span data-contrast=\"auto\">The model cannot be\u00a0the\u00a0accountable party. Accountability stays with the organization deploying it.<\/span><\/i><\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71afff\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<p data-start=\"137\" data-end=\"542\">This isn&#8217;t just an internal best practice. It&#8217;s becoming an industry standard. As <a class=\"decorated-link\" href=\"https:\/\/nhimg.org\/faq\/who-is-accountable-for-fraud-risk-decisions-when-ai-is-used-in-detection-and-pre\/\" target=\"_new\" rel=\"noopener\" data-start=\"219\" data-end=\"352\">NHI Management Group notes<\/a>, accountability does not shift to the model, platform, or data science team when AI enters fraud detection. It remains with the organization using the system to make or influence decisions. Regulators increasingly expect evidence of this accountability, not just a policy statement. Frameworks like NIST&#8217;s AI Risk Management Framework and NIST SP 800-53 now treat named ownership and audit logging as core governance controls. They are no longer optional extras. These frameworks also require separation of duties. Model builders, case reviewers, and approvers should not control the entire decision chain alone. Put simply, &#8220;who is accountable&#8221; is no longer just a philosophical question. It is becoming a compliance requirement.<\/p>\n<p data-start=\"1087\" data-end=\"1331\">This gap often widens when AI systems move beyond passive scoring and start making autonomous decisions. The progression goes from &#8220;flag this transaction&#8221; to &#8220;hold this payout&#8221; and increasingly to &#8220;close this account&#8221; without human involvement. We explore this shift in <a class=\"decorated-link\" href=\"https:\/\/smartdev.com\/fr\/your-ai-agent-has-access-but-does-it-have-any-identity\/\" target=\"_new\" rel=\"noopener\" data-start=\"1358\" data-end=\"1494\">Your AI Agent Has Access, But Does It Have Any Identity?<\/a>. The article examines what happens when an AI agent can take action without a clearly attributable identity. When an action cannot be traced back to a specific person, policy, or approval, organizations lose the ability to determine who &#8211; or what &#8211; is accountable.<\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71b155\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<p class=\"PDq2pG_selectionAnchorContainer\" data-start=\"66\" data-end=\"270\">The same gap tends to open up during the transition from pilot to production, and for a similar reason. Pilots run under close supervision, with small case volumes and engineers monitoring every output. Production is different. It runs at scale, with less supervision, and affects real customers. As <a class=\"decorated-link\" href=\"https:\/\/smartdev.com\/fr\/from-ai-pilot-to-controlled-production-what-your-team-needs-to-own-and-what-you-can-outsource\/\" target=\"_new\" rel=\"noopener\" data-start=\"372\" data-end=\"529\">From AI Pilot to Controlled Production<\/a> explains, deciding what your team must own internally versus what it can safely outsource is a critical governance decision. It determines who becomes accountable later. Teams that skip this step while rushing to scale often discover the accountability gap only after a customer, journalist, or regulator asks the question first.<\/p>\n<p data-start=\"872\" data-end=\"1074\">This principle applies especially to high-consequence actions: account suspension, transaction rejection, payout freezing, seller termination, and escalating suspicious activity reports to regulators. Each action needs a clearly defined decision owner &#8211; a named role, not a team. It should never be assumed that \u201cthe system decided.\u201d Ownership at the department level dilutes accountability just as much as blaming \u201cthe algorithm.\u201d<\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71b287\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<h3 aria-level=\"3\"><span class=\"ez-toc-section\" id=\"Human-in-the-Loop_Is_Not_the_Same_as_Human_Accountability\"><\/span><b><span data-contrast=\"none\">Human-in-the-Loop Is\u00a0Not the Same as\u00a0Human Accountability<\/span><\/b><span data-ccp-props=\"{&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;335559738&quot;:160,&quot;335559739&quot;:80}\">\u00a0<\/span><span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><span data-contrast=\"auto\">This is the second major argument of this piece,\u00a0and also\u00a0the one most organizations get wrong\u00a0&#8211;\u00a0often while believing\u00a0they&#8217;ve\u00a0already solved it, simply because a human sits somewhere in the workflow.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">A process that includes a human reviewer\u00a0doesn&#8217;t\u00a0automatically guarantee meaningful oversight. Consider a common flow:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p style=\"text-align: center;\"><i><span data-contrast=\"auto\">AI flags \u2192 Human approves \u2192 System acts<\/span><\/i><span data-ccp-props=\"{&quot;335551550&quot;:2,&quot;335551620&quot;:2}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">At first glance, it\u00a0sounds safe. It even satisfies most internal audit checklists, since &#8220;human-in-the-loop&#8221; appears as a control on paper. Look closely at the quality of that &#8220;human approves&#8221; step, though, and the picture changes. Suppose the reviewer:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"6\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Doesn&#8217;t\u00a0understand the evidence behind the alert, and simply sees a risk score with no supporting detail,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"11\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Can&#8217;t\u00a0see the model&#8217;s confidence level, so an 82%-confidence flag looks identical to a 51%-confidence flag,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"12\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Has no real authority to override the decision, because the workflow only lets them confirm or pass the case along, or<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"13\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Juggles too many alerts to think carefully about any single\u00a0one, and\u00a0works under a quota that rewards speed over judgment.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<p class=\"PDq2pG_selectionAnchorContainer\" data-start=\"71\" data-end=\"391\">In any of these cases, the \u201chuman approval\u201d step becomes purely procedural. It turns into procedural theatre &#8211; a ritual that creates the appearance of oversight without real substance. Worse, it can make things more dangerous by giving the organization false confidence that a person has already reviewed the decision.<\/p>\n<p data-start=\"396\" data-end=\"678\">This isn\u2019t a hypothetical risk. Alert fatigue remains a well-documented problem across fraud and AML operations. When most alerts become false positives, <a class=\"decorated-link\" href=\"https:\/\/youverify.co\/blog\/howreduce-aml-false-positives\" target=\"_new\" rel=\"noopener\" data-start=\"550\" data-end=\"675\">analysts\u2019 attention to genuinely suspicious cases degrades sharply<\/a>. Picture a team processing 500 alerts a day with a high false-positive rate. They are searching for a few dozen real cases buried under hundreds of noise cases. Under sustained pressure, reviewers develop shortcuts as a survival mechanism. They rely on surface patterns, clear familiar cases quickly, and reserve deeper scrutiny for only the most unusual alerts. This is a rational response to an unreasonable workload. However, it quietly weakens the oversight the process was designed to provide. Under these conditions, \u201ca human signed off\u201d no longer represents meaningful oversight. It is simply a signature passing through.<\/p>\n<p><span data-contrast=\"auto\">So\u00a0what does meaningful human oversight actually require? Three conditions need to hold at once, and all three\u00a0have to\u00a0be present simultaneously\u00a0&#8211;\u00a0having two out of three still leaves a gap wide enough for bad decisions to slip through unchallenged.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"14\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Information\u00a0&#8211;\u00a0enough evidence to actually understand and challenge the AI&#8217;s conclusion.\u00a0This means seeing not just a score, but the features and context that produced it: which signals fired, how they compare to the customer&#8217;s normal behavior, and what similar past cases resolved to.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"15\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Authority\u00a0&#8211;\u00a0the actual power to override or escalate, not just the ability to click &#8220;approve.&#8221; A reviewer who can only confirm the system&#8217;s recommendation, and never reverse it,\u00a0isn&#8217;t\u00a0exercising authority;\u00a0they&#8217;re\u00a0rubber-stamping a decision the system already made.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"16\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Capacity\u00a0&#8211;\u00a0enough time and a reasonable workload to genuinely exercise judgment, rather than rubber-stamping to hit a quota. An organization that measures reviewer performance purely by throughput is quietly\u00a0optimizing away\u00a0the judgment it claims to want.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71b4e2\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<p><span class=\"TextRun SCXW166615052 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW166615052 BCX8\">Miss any one of these three conditions, and human-in-the-loop risks becoming a thin layer of legitimacy wrapped around a decision the machine still effectively makes on its own. Compliance teams have lived through the exact same shift. As\u00a0<\/span><\/span><a class=\"Hyperlink SCXW166615052 BCX8\" href=\"https:\/\/smartdev.com\/fr\/from-processing-alerts-to-making-decisions-how-ai-is-redefining-compliance-teams\/\" target=\"_blank\" rel=\"noreferrer noopener\"><span class=\"TextRun Underlined SCXW166615052 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW166615052 BCX8\" data-ccp-charstyle=\"Hyperlink\">From Processing Alerts to Making Decisions<\/span><\/span><\/a><span class=\"TextRun SCXW166615052 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW166615052 BCX8\">\u00a0<\/span><span class=\"NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW166615052 BCX8\">shows<\/span><span class=\"NormalTextRun SCXW166615052 BCX8\">, AI only delivers real value once teams stop treating it as an alert generator and start giving reviewers the context and the authority to actually decide.<\/span><\/span><span class=\"EOP Selected SCXW166615052 BCX8\" data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Given all this, a risk-based escalation model offers a more sensible path forward than either full automation or reviewing every\u00a0single case\u00a0by hand:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"17\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Low risk\u00a0+ high confidence \u2192 safe to automate,\u00a0freeing\u00a0reviewers to focus where judgment\u00a0actually adds\u00a0value.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"18\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">Ambiguous cases &#8211; moderate confidence, conflicting signals, or unfamiliar patterns \u2192 mandatory human review, with enough context supplied to make that review meaningful rather than nominal.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"19\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><span data-contrast=\"auto\">High-impact decisions (account closure, seller suspension, freezing large sums) \u2192 mandatory human oversight, no exceptions, regardless of how confident the model is. The size of the potential harm, not the model&#8217;s confidence score, should set the bar for when a human must be involved.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">It&#8217;s\u00a0also worth distinguishing between two models of oversight that often get treated as interchangeable.\u00a0<\/span><a href=\"https:\/\/nhimg.org\/faq\/who-is-accountable-for-fraud-risk-decisions-when-ai-is-used-in-detection-and-pre\/\"><span data-contrast=\"none\">NHI Management Group draws a useful line<\/span><\/a><span data-contrast=\"auto\">\u00a0between &#8220;human-in-the-loop,&#8221; where a person reviews and approves before the system acts, and &#8220;human-on-the-loop,&#8221; where the system acts\u00a0first\u00a0and a person monitors and reviews after the fact. Neither model is inherently better; the right choice depends on the organization&#8217;s risk appetite, how much harm a wrong action could cause, and how mature its audit trail is. What both models share, though, is the same non-negotiable requirement: someone has to be positioned\u00a0&#8211;\u00a0with the information, authority, and capacity described above\u00a0&#8211;\u00a0to actually catch and correct a bad decision, whether that happens before or after the system acts.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-40724 size-full\" src=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_56_04-AM.png\" alt=\"\" width=\"1536\" height=\"1024\" srcset=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_56_04-AM.png 1536w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_56_04-AM-300x200.png 300w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_56_04-AM-1024x683.png 1024w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_56_04-AM-768x512.png 768w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-10_56_04-AM-18x12.png 18w\" sizes=\"auto, (max-width: 1536px) 100vw, 1536px\" \/><\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71b673\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<h3><span class=\"ez-toc-section\" id=\"What_Accountable_AI_Fraud_Detection_Should_Look_Like\"><\/span><span class=\"TextRun SCXW63168875 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW63168875 BCX8\" data-ccp-parastyle=\"heading 3\">What Accountable AI Fraud Detection Should Look Like<\/span><\/span><span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><span data-contrast=\"auto\">Rather than a long checklist, four core pillars matter most\u00a0&#8211;\u00a0and each one addresses a specific failure mode already covered above.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"20\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559683&quot;:0,&quot;335559684&quot;:-2,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><b><span data-contrast=\"auto\">Clear ownership<\/span><\/b><span data-contrast=\"auto\">.\u00a0<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335551550&quot;:0,&quot;335551620&quot;:0,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">The question\u00a0shouldn&#8217;t\u00a0stop at &#8220;who owns the model?&#8221; It needs to go further:\u00a0<\/span><b><span data-contrast=\"auto\">who owns the decision the model triggers?<\/span><\/b><span data-contrast=\"auto\">\u00a0That distinction separates technical responsibility from decision-making responsibility, and it directly answers the accountability gap described earlier. In practice, this means naming a specific role &#8211; not a department &#8211; for every category of consequential action a fraud system can take, with the authority to change thresholds, pause automation, or demand a review. Without a named owner, &#8220;the model did it&#8221; quietly becomes an acceptable answer inside the organization, even though\u00a0it&#8217;s\u00a0never acceptable to a regulator or a harmed customer.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335551550&quot;:0,&quot;335551620&quot;:0,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"21\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559683&quot;:0,&quot;335559684&quot;:-2,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><b><span data-contrast=\"auto\">Decision traceability<\/span><\/b><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335551550&quot;:0,&quot;335551620&quot;:0,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">\u00a0An organization must be able to reconstruct the full path leading to any action: what data existed at the time, which model version the team used, which signals fired, what recommendation the system produced, whether a human intervened, and what final action followed. This record needs to exist at the level of the individual case, not just in aggregate &#8211; a monthly accuracy report says nothing about why one specific\u00a0customer&#8217;s\u00a0account got frozen last Tuesday. Without case-level traceability, an organization\u00a0can&#8217;t\u00a0learn from its mistakes,\u00a0can&#8217;t\u00a0tell a one-off error from a systemic pattern, and\u00a0can&#8217;t\u00a0demonstrate\u00a0it acted reasonably when challenged. This principle underlies a defensible\u00a0<\/span><a href=\"https:\/\/smartdev.com\/fr\/ai-compliance-audit-trail\/\"><span data-contrast=\"none\">AI compliance audit trail<\/span><\/a><span data-contrast=\"auto\">: every AI-assisted decision needs a\u00a0reconstructable\u00a0record that survives long after the case itself is closed.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335551550&quot;:0,&quot;335551620&quot;:0,&quot;335559685&quot;:0,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"22\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559683&quot;:0,&quot;335559684&quot;:-2,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><b><span data-contrast=\"auto\">Risk-based human intervention<\/span><\/b><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335551550&quot;:0,&quot;335551620&quot;:0,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">Teams should base the decision to bring in a human on concrete factors: the model&#8217;s confidence level, the potential financial exposure, the customer impact, and the regulatory sensitivity of the case. Crucially, these thresholds\u00a0shouldn&#8217;t\u00a0be set once and left alone &#8211; fraud patterns evolve, and a threshold calibrated for last year&#8217;s fraud mix can quietly become\u00a0miscalibrated\u00a0for this year&#8217;s. A regular\u00a0review\u00a0cadence keeps human intervention pointed at the cases that actually carry the most risk today.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335551550&quot;:0,&quot;335551620&quot;:0,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<ul>\n<li aria-setsize=\"-1\" data-leveltext=\"\uf0b7\" data-font=\"Symbol\" data-listid=\"23\" data-list-defn-props=\"{&quot;335552541&quot;:1,&quot;335559683&quot;:0,&quot;335559684&quot;:-2,&quot;335559685&quot;:720,&quot;335559991&quot;:360,&quot;469769226&quot;:&quot;Symbol&quot;,&quot;469769242&quot;:&#091;8226&#093;,&quot;469777803&quot;:&quot;left&quot;,&quot;469777804&quot;:&quot;\uf0b7&quot;,&quot;469777815&quot;:&quot;hybridMultilevel&quot;}\" data-aria-posinset=\"1\" data-aria-level=\"1\"><b><span data-contrast=\"auto\">Continuous feedback<\/span><\/b><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335551550&quot;:0,&quot;335551620&quot;:0,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">Confirmed fraud cases, false positives, human overrides, and customer disputes should all feed back into improving the model, the thresholds, the business rules, and the escalation logic. A well-designed feedback loop treats every human override as a signal worth investigating: if analysts routinely override the model on the same type of case,\u00a0that&#8217;s\u00a0a\u00a0sign\u00a0the threshold &#8211; not the analyst &#8211; needs adjusting. Without this loop, a system will keep repeating the same false positives and false negatives indefinitely, since nothing in the process ever flags that something went wrong.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335551550&quot;:0,&quot;335551620&quot;:0,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-40726 size-full\" src=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_02_27-AM.png\" alt=\"\" width=\"1536\" height=\"1024\" srcset=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_02_27-AM.png 1536w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_02_27-AM-300x200.png 300w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_02_27-AM-1024x683.png 1024w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_02_27-AM-768x512.png 768w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_02_27-AM-18x12.png 18w\" sizes=\"auto, (max-width: 1536px) 100vw, 1536px\" \/><\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71b889\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<h3 aria-level=\"3\"><span class=\"ez-toc-section\" id=\"The_Better_Goal_AI-Assisted_Fraud_Judgment_Not_Autonomous_Fraud_Decisions\"><\/span><b><span data-contrast=\"none\">The Better Goal: AI-Assisted Fraud Judgment, Not Autonomous Fraud Decisions<\/span><\/b><span data-ccp-props=\"{&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;335559738&quot;:160,&quot;335559739&quot;:80}\">\u00a0<\/span><span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><span data-contrast=\"auto\">This is the strategic takeaway of the whole piece.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">AI excels at pattern recognition, anomaly detection, prioritization, and aggregating evidence at scale. Humans, meanwhile, excel at the things AI still struggles\u00a0with:\u00a0understanding context, navigating ambiguity, weighing proportionality between the level of suspicion and the severity of the action taken, and exercising final judgment.<\/span><\/p>\n<p><span data-contrast=\"auto\">Given these complementary strengths, the operating model worth aiming for looks like this:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p style=\"text-align: center;\"><i><span data-contrast=\"auto\">AI detects \u2192 AI assembles evidence \u2192 Humans judge consequential cases \u2192 Outcomes\u00a0feed back\u00a0to improve the system<\/span><\/i><span data-ccp-props=\"{&quot;335551550&quot;:2,&quot;335551620&quot;:2}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Not this:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p style=\"text-align: center;\"><i><span data-contrast=\"auto\">AI detects \u2192 AI decides \u2192 Human signs off<\/span><\/i><span data-ccp-props=\"{&quot;335551550&quot;:2,&quot;335551620&quot;:2}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-40725 size-full\" src=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_00_14-AM.png\" alt=\"\" width=\"1536\" height=\"1024\" srcset=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_00_14-AM.png 1536w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_00_14-AM-300x200.png 300w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_00_14-AM-1024x683.png 1024w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_00_14-AM-768x512.png 768w, https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_00_14-AM-18x12.png 18w\" sizes=\"auto, (max-width: 1536px) 100vw, 1536px\" \/><\/span><\/p>\n<p><span data-contrast=\"auto\">It&#8217;s\u00a0also worth noting that teams\u00a0shouldn&#8217;t\u00a0measure success by model accuracy alone. Several other metrics deserve equal attention: the customer impact of false positives, the investigation workload the system\u00a0creates,\u00a0the rate at which humans override AI decisions, the friction imposed on customers, the losses actually prevented, and the consistency of decisions over time. After all, a model can post high accuracy on paper while it quietly burns out analysts through alert\u00a0fatigue, or\u00a0drives away a third of the customers it wrongly declines. Neither outcome counts as success by any meaningful business measure.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>\n\t\t<div id=\"fws_6aaab7d71b9f0\"  data-column-margin=\"default\" data-midnight=\"dark\"  class=\"wpb_row vc_row-fluid vc_row\"  style=\"padding-top: 0px; padding-bottom: 0px; \"><div class=\"row-bg-wrap\" data-bg-animation=\"none\" data-bg-animation-delay=\"\" data-bg-overlay=\"false\"><div class=\"inner-wrap row-bg-layer\" ><div class=\"row-bg viewport-desktop\"  style=\"\"><\/div><\/div><\/div><div class=\"row_col_wrap_12 col span_12 dark left\">\n\t<div  class=\"vc_col-sm-12 wpb_column column_container vc_column_container col no-extra-padding inherit_tablet inherit_phone flex_gap_desktop_10px\"  data-padding-pos=\"all\" data-has-bg-color=\"false\" data-bg-color=\"\" data-bg-opacity=\"1\" data-animation=\"\" data-delay=\"0\" >\n\t\t<div class=\"vc_column-inner\" >\n\t\t\t<div class=\"wpb_wrapper\">\n\t\t\t\t\n<div class=\"wpb_text_column wpb_content_element\" >\n\t<h3 dir=\"ltr\"><span class=\"ez-toc-section\" id=\"Conclusion_%E2%80%93_Accountability_Cannot_Be_Outsourced_to_an_Algorithm\"><\/span>Conclusion &#8211; Accountability Cannot Be Outsourced to an Algorithm<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p dir=\"ltr\">AI can process more data than humans, spot patterns faster, and prioritize suspicious cases more effectively. Even so, it still operates inside a decision system that humans designed.<\/p>\n<p dir=\"ltr\">Humans set the thresholds. The organization approves the rules. The business chooses the boundaries of automation; the model never decides them on its own.<\/p>\n<p dir=\"ltr\">So when AI gets fraud detection wrong, the answer can&#8217;t simply be &#8220;the model made a mistake.&#8221; Ultimately, accountability has to sit with the organization, and with the people who designed, governed, and operated that system.<\/p>\n<p dir=\"ltr\">AI can generate the fraud signal. It can even recommend the action. But accountability for the decision must remain human.<\/p>\n<p dir=\"ltr\">Getting this right rarely comes down to picking a better model. It comes down to designing the governance around it: clear ownership, traceable decisions, and human review that actually has teeth. That&#8217;s the kind of work our team at SmartDev spends most of its time on with fraud and compliance teams building AI into consequential decisions \u2014 less about shipping a smarter classifier, more about making sure the system it feeds into can stand behind every decision it makes. If you&#8217;re wrestling with where to draw that line in your own fraud stack, <a href=\"https:\/\/smartdev.com\/fr\/contact-us\/\">we&#8217;re happy to talk it through<\/a>.<\/p>\n<\/div>\n\n\n\n\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n<\/div><\/div>","protected":false},"excerpt":{"rendered":"TL;DR AI generates risk signals, not decisions - organizations choose what those signals are allowed...","protected":false},"author":46,"featured_media":40727,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"inline_featured_image":false,"footnotes":""},"categories":[236,518,74,247],"tags":[71,66,638],"class_list":["post-40716","post","type-post","status-publish","format-standard","has-post-thumbnail","category-ai-adoption","category-nora","category-services","category-workflow-automation","tag-ai-adoption","tag-smartdev","tag-workflow-automation"],"acf":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v28.5 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>When AI Gets Fraud Detection Wrong: Who Is Accountable? | SmartDev<\/title>\n<meta name=\"description\" content=\"AI can flag fraud, but it can&#039;t own the decision. Here&#039;s why accountability for AI-driven fraud detection must always stay with humans, not the model.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"When AI Gets Fraud Detection Wrong: Who Is Accountable? | SmartDev\" \/>\n<meta property=\"og:description\" content=\"AI can flag fraud, but it can&#039;t own the decision. Here&#039;s why accountability for AI-driven fraud detection must always stay with humans, not the model.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/\" \/>\n<meta property=\"og:site_name\" content=\"SmartDev\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.youtube.com\/@smartdevllc\" \/>\n<meta property=\"article:published_time\" content=\"2026-09-16T06:46:40+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_30_11-AM.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1774\" \/>\n\t<meta property=\"og:image:height\" content=\"887\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"Uyen Nguyen\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@smartdevllc\" \/>\n<meta name=\"twitter:site\" content=\"@smartdevllc\" \/>\n<meta name=\"twitter:label1\" content=\"\u00c9crit par\" \/>\n\t<meta name=\"twitter:data1\" content=\"Uyen Nguyen\" \/>\n\t<meta name=\"twitter:label2\" content=\"Dur\u00e9e de lecture estim\u00e9e\" \/>\n\t<meta name=\"twitter:data2\" content=\"17 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/when-ai-gets-fraud-detection-wrong-who-is-accountable\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/when-ai-gets-fraud-detection-wrong-who-is-accountable\\\/\"},\"author\":{\"name\":\"Uyen Nguyen\",\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/#\\\/schema\\\/person\\\/f7a8201f9f8bc8a852880192ff658251\"},\"headline\":\"When AI Gets Fraud Detection Wrong: Who Is Accountable?\",\"datePublished\":\"2026-09-16T06:46:40+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/when-ai-gets-fraud-detection-wrong-who-is-accountable\\\/\"},\"wordCount\":6195,\"publisher\":{\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/when-ai-gets-fraud-detection-wrong-who-is-accountable\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/smartdev.com\\\/wp-content\\\/uploads\\\/2026\\\/09\\\/ChatGPT-Image-Sep-10-2026-11_30_11-AM.png\",\"keywords\":[\"AI Adoption\",\"SmartDev\",\"Workflow Automation\"],\"articleSection\":[\"AI Adoption\",\"NORA\",\"Services\",\"Workflow Automation\"],\"inLanguage\":\"fr-FR\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/when-ai-gets-fraud-detection-wrong-who-is-accountable\\\/\",\"url\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/when-ai-gets-fraud-detection-wrong-who-is-accountable\\\/\",\"name\":\"When AI Gets Fraud Detection Wrong: Who Is Accountable? | SmartDev\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/when-ai-gets-fraud-detection-wrong-who-is-accountable\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/when-ai-gets-fraud-detection-wrong-who-is-accountable\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/smartdev.com\\\/wp-content\\\/uploads\\\/2026\\\/09\\\/ChatGPT-Image-Sep-10-2026-11_30_11-AM.png\",\"datePublished\":\"2026-09-16T06:46:40+00:00\",\"description\":\"AI can flag fraud, but it can't own the decision. Here's why accountability for AI-driven fraud detection must always stay with humans, not the model.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/when-ai-gets-fraud-detection-wrong-who-is-accountable\\\/#breadcrumb\"},\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/smartdev.com\\\/fr\\\/when-ai-gets-fraud-detection-wrong-who-is-accountable\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/when-ai-gets-fraud-detection-wrong-who-is-accountable\\\/#primaryimage\",\"url\":\"https:\\\/\\\/smartdev.com\\\/wp-content\\\/uploads\\\/2026\\\/09\\\/ChatGPT-Image-Sep-10-2026-11_30_11-AM.png\",\"contentUrl\":\"https:\\\/\\\/smartdev.com\\\/wp-content\\\/uploads\\\/2026\\\/09\\\/ChatGPT-Image-Sep-10-2026-11_30_11-AM.png\",\"width\":1774,\"height\":887},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/when-ai-gets-fraud-detection-wrong-who-is-accountable\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/smartdev.com\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"When AI Gets Fraud Detection Wrong: Who Is Accountable?\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/#website\",\"url\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/\",\"name\":\"SmartDev\",\"description\":\"Al Powered Software Development\",\"publisher\":{\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/#organization\"},\"alternateName\":\"SmartDev\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"fr-FR\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/#organization\",\"name\":\"SmartDev\",\"alternateName\":\"SmartDev\",\"url\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/smartdev.com\\\/wp-content\\\/uploads\\\/2025\\\/04\\\/SMD-Logo-New-Main-scaled.png\",\"contentUrl\":\"https:\\\/\\\/smartdev.com\\\/wp-content\\\/uploads\\\/2025\\\/04\\\/SMD-Logo-New-Main-scaled.png\",\"width\":2560,\"height\":550,\"caption\":\"SmartDev\"},\"image\":{\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/www.youtube.com\\\/@smartdevllc\",\"https:\\\/\\\/x.com\\\/smartdevllc\",\"https:\\\/\\\/www.linkedin.com\\\/company\\\/4873071\\\/\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/#\\\/schema\\\/person\\\/f7a8201f9f8bc8a852880192ff658251\",\"name\":\"Uyen Nguyen\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/36c4a3d2a7aef0d7fa216ac82ad5e150f2560bf5cc5167166ff0846f0117e4d7?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/36c4a3d2a7aef0d7fa216ac82ad5e150f2560bf5cc5167166ff0846f0117e4d7?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/36c4a3d2a7aef0d7fa216ac82ad5e150f2560bf5cc5167166ff0846f0117e4d7?s=96&d=mm&r=g\",\"caption\":\"Uyen Nguyen\"},\"description\":\"She is a marketing professional with a deep passion for leveraging digital technologies and AI to enhance marketing effectiveness. With extensive knowledge in AI implementation and hands-on experience at SmartDev, she is committed to providing valuable insights and perspectives on AI integration across diverse industries, aiming to drive operational excellence and business growth.\",\"url\":\"https:\\\/\\\/smartdev.com\\\/fr\\\/author\\\/uyen-nguyentranphuong\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"When AI Gets Fraud Detection Wrong: Who Is Accountable? | SmartDev","description":"AI can flag fraud, but it can't own the decision. Here's why accountability for AI-driven fraud detection must always stay with humans, not the model.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/","og_locale":"fr_FR","og_type":"article","og_title":"When AI Gets Fraud Detection Wrong: Who Is Accountable? | SmartDev","og_description":"AI can flag fraud, but it can't own the decision. Here's why accountability for AI-driven fraud detection must always stay with humans, not the model.","og_url":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/","og_site_name":"SmartDev","article_publisher":"https:\/\/www.youtube.com\/@smartdevllc","article_published_time":"2026-09-16T06:46:40+00:00","og_image":[{"width":1774,"height":887,"url":"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_30_11-AM.png","type":"image\/png"}],"author":"Uyen Nguyen","twitter_card":"summary_large_image","twitter_creator":"@smartdevllc","twitter_site":"@smartdevllc","twitter_misc":{"\u00c9crit par":"Uyen Nguyen","Dur\u00e9e de lecture estim\u00e9e":"17 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/#article","isPartOf":{"@id":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/"},"author":{"name":"Uyen Nguyen","@id":"https:\/\/smartdev.com\/fr\/#\/schema\/person\/f7a8201f9f8bc8a852880192ff658251"},"headline":"When AI Gets Fraud Detection Wrong: Who Is Accountable?","datePublished":"2026-09-16T06:46:40+00:00","mainEntityOfPage":{"@id":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/"},"wordCount":6195,"publisher":{"@id":"https:\/\/smartdev.com\/fr\/#organization"},"image":{"@id":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/#primaryimage"},"thumbnailUrl":"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_30_11-AM.png","keywords":["AI Adoption","SmartDev","Workflow Automation"],"articleSection":["AI Adoption","NORA","Services","Workflow Automation"],"inLanguage":"fr-FR"},{"@type":"WebPage","@id":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/","url":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/","name":"When AI Gets Fraud Detection Wrong: Who Is Accountable? | SmartDev","isPartOf":{"@id":"https:\/\/smartdev.com\/fr\/#website"},"primaryImageOfPage":{"@id":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/#primaryimage"},"image":{"@id":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/#primaryimage"},"thumbnailUrl":"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_30_11-AM.png","datePublished":"2026-09-16T06:46:40+00:00","description":"AI can flag fraud, but it can't own the decision. Here's why accountability for AI-driven fraud detection must always stay with humans, not the model.","breadcrumb":{"@id":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/#breadcrumb"},"inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/"]}]},{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/#primaryimage","url":"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_30_11-AM.png","contentUrl":"https:\/\/smartdev.com\/wp-content\/uploads\/2026\/09\/ChatGPT-Image-Sep-10-2026-11_30_11-AM.png","width":1774,"height":887},{"@type":"BreadcrumbList","@id":"https:\/\/smartdev.com\/fr\/when-ai-gets-fraud-detection-wrong-who-is-accountable\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/smartdev.com\/"},{"@type":"ListItem","position":2,"name":"When AI Gets Fraud Detection Wrong: Who Is Accountable?"}]},{"@type":"WebSite","@id":"https:\/\/smartdev.com\/fr\/#website","url":"https:\/\/smartdev.com\/fr\/","name":"SmartDev","description":"D\u00e9veloppement de logiciels aliment\u00e9 par l&#039;IA","publisher":{"@id":"https:\/\/smartdev.com\/fr\/#organization"},"alternateName":"SmartDev","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/smartdev.com\/fr\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"fr-FR"},{"@type":"Organization","@id":"https:\/\/smartdev.com\/fr\/#organization","name":"SmartDev","alternateName":"SmartDev","url":"https:\/\/smartdev.com\/fr\/","logo":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/smartdev.com\/fr\/#\/schema\/logo\/image\/","url":"https:\/\/smartdev.com\/wp-content\/uploads\/2025\/04\/SMD-Logo-New-Main-scaled.png","contentUrl":"https:\/\/smartdev.com\/wp-content\/uploads\/2025\/04\/SMD-Logo-New-Main-scaled.png","width":2560,"height":550,"caption":"SmartDev"},"image":{"@id":"https:\/\/smartdev.com\/fr\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.youtube.com\/@smartdevllc","https:\/\/x.com\/smartdevllc","https:\/\/www.linkedin.com\/company\/4873071\/"]},{"@type":"Person","@id":"https:\/\/smartdev.com\/fr\/#\/schema\/person\/f7a8201f9f8bc8a852880192ff658251","name":"Uyen Nguyen","image":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/secure.gravatar.com\/avatar\/36c4a3d2a7aef0d7fa216ac82ad5e150f2560bf5cc5167166ff0846f0117e4d7?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/36c4a3d2a7aef0d7fa216ac82ad5e150f2560bf5cc5167166ff0846f0117e4d7?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/36c4a3d2a7aef0d7fa216ac82ad5e150f2560bf5cc5167166ff0846f0117e4d7?s=96&d=mm&r=g","caption":"Uyen Nguyen"},"description":"She is a marketing professional with a deep passion for leveraging digital technologies and AI to enhance marketing effectiveness. With extensive knowledge in AI implementation and hands-on experience at SmartDev, she is committed to providing valuable insights and perspectives on AI integration across diverse industries, aiming to drive operational excellence and business growth.","url":"https:\/\/smartdev.com\/fr\/author\/uyen-nguyentranphuong\/"}]}},"_links":{"self":[{"href":"https:\/\/smartdev.com\/fr\/wp-json\/wp\/v2\/posts\/40716","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/smartdev.com\/fr\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/smartdev.com\/fr\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/smartdev.com\/fr\/wp-json\/wp\/v2\/users\/46"}],"replies":[{"embeddable":true,"href":"https:\/\/smartdev.com\/fr\/wp-json\/wp\/v2\/comments?post=40716"}],"version-history":[{"count":1,"href":"https:\/\/smartdev.com\/fr\/wp-json\/wp\/v2\/posts\/40716\/revisions"}],"predecessor-version":[{"id":40728,"href":"https:\/\/smartdev.com\/fr\/wp-json\/wp\/v2\/posts\/40716\/revisions\/40728"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/smartdev.com\/fr\/wp-json\/wp\/v2\/media\/40727"}],"wp:attachment":[{"href":"https:\/\/smartdev.com\/fr\/wp-json\/wp\/v2\/media?parent=40716"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/smartdev.com\/fr\/wp-json\/wp\/v2\/categories?post=40716"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/smartdev.com\/fr\/wp-json\/wp\/v2\/tags?post=40716"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}