{"id":8117,"date":"2026-10-01T11:43:32","date_gmt":"2026-10-01T15:43:32","guid":{"rendered":"https:\/\/verito.com\/blog\/?p=8117"},"modified":"2026-09-08T14:09:11","modified_gmt":"2026-09-08T18:09:11","slug":"shadow-ai-client-data-chatgpt","status":"publish","type":"post","link":"https:\/\/verito.com\/blog\/shadow-ai-client-data-chatgpt\/","title":{"rendered":"What Do You Actually Do When Staff Are Pasting Client Data Into ChatGPT?"},"content":{"rendered":"\n<p><strong>It\u2019s likely that someone at your firm has already pasted a client\u2019s information into ChatGPT, Claude, or Gemini to save time on a draft, a summary, or a tricky question. Not out of carelessness. Out of the same instinct that makes any of us reach for a faster tool. That\u2019s shadow AI: AI use your firm never approved, sitting outside any policy you\u2019ve written, on a device you already manage. The fix is knowing what\u2019s actually being used, deciding what\u2019s approved, and backing that decision with something more than good intentions.<\/strong><\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p>Key takeaways<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Shadow AI isn\u2019t a hypothetical. Staff are very likely already using consumer AI tools with client data, whether or not the firm has approved it.<\/li>\n\n\n\n<li>Pasting client tax data into a public AI tool touches the same territory as Section 7216 and the FTC Safeguards Rule, covered in depth elsewhere on this blog.<\/li>\n\n\n\n<li>Finding out what\u2019s already in use starts with a short, non-punitive ask, not a surprise audit.<\/li>\n\n\n\n<li>A usable tool-vetting standard has three questions: where does the data live, is the vendor SOC 2 compliant, and what do the retention and deletion terms actually say.<\/li>\n\n\n\n<li>A written policy without technical backing relies entirely on someone remembering the rule at 4pm during a deadline crunch.<\/li>\n<\/ul>\n<\/blockquote>\n\n\n\n<div class=\"wp-block-rank-math-toc-block\" id=\"rank-math-toc\"><h2 id=\"table-of-contents\">Table of Contents<\/h2><nav><ul><li><a href=\"#why-shadow-ai-isnt-hypothetical-for-your-firm\">Why shadow AI isn\u2019t hypothetical for your firm<\/a><\/li><li><a href=\"#whats-actually-at-stake\">What\u2019s actually at stake<\/a><\/li><li><a href=\"#how-do-you-find-out-whats-already-being-used\">How do you find out what\u2019s already being used?<\/a><\/li><li><a href=\"#what-does-an-approved-tool-list-actually-look-like\">What does an approved-tool list actually look like?<\/a><\/li><li><a href=\"#where-policy-needs-technical-backing\">Where policy needs technical backing<\/a><\/li><\/ul><\/nav><\/div>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"why-shadow-ai-isnt-hypothetical-for-your-firm\"><strong>Why shadow AI isn\u2019t hypothetical for your firm<\/strong><\/h2>\n\n\n\n<p><strong>A five-person practice doesn\u2019t have a compliance team standing between a new AI tool and a staff member\u2019s next deadline. A large firm\u2019s IT and legal departments vet a tool before anyone\u2019s allowed near it with client data. A small firm usually finds out after someone\u2019s already been using it for months.<\/strong><\/p>\n\n\n\n<p>It\u2019s a structural gap: the review that happens automatically at a large firm has to happen deliberately at a small one, and deliberate is easy to skip when nothing\u2019s gone wrong yet. The person who pasted a client\u2019s return into a chatbot to draft a summary wasn\u2019t trying to create an incident. They were trying to save twenty minutes on a Tuesday.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"whats-actually-at-stake\"><strong>What\u2019s actually at stake<\/strong><\/h2>\n\n\n\n<p><strong>Once client data goes into a consumer AI tool, the firm has lost visibility into where it went, who can access it, and how long it\u2019s kept, and that\u2019s before getting into the specific legal exposure.<\/strong><\/p>\n\n\n\n<p>Section 7216 governs when a firm can disclose or use a client\u2019s return information, and the <a href=\"https:\/\/verito.com\/blog\/irs-circular-230-ai-guidance-tax-practitioners\/\" target=\"_blank\" rel=\"dofollow\" >Circular 230 duties the IRS has attached to AI use<\/a> add a separate layer of professional obligation on top of that, including an open question about whether AI use itself needs to be disclosed to the client. Both are worth reading in full rather than summarized here. What matters for this piece is simpler: the exposure isn\u2019t theoretical, and it doesn\u2019t wait for the firm to have a policy before it applies.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"how-do-you-find-out-whats-already-being-used\"><strong>How do you find out what\u2019s already being used?<\/strong><\/h2>\n\n\n\n<p><strong>Start by asking, not auditing. A short, explicitly non-punitive survey, five questions or fewer, catches more real usage than any technical scan, because people don\u2019t hide a productivity shortcut when nobody\u2019s in trouble for it.<\/strong><\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"576\" src=\"https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-survey-1024x576.jpg\" alt=\"Staff AI usage survey cinematic visual | Verito\" class=\"wp-image-8138\" srcset=\"https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-survey-1024x576.jpg 1024w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-survey-300x169.jpg 300w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-survey-768x432.jpg 768w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-survey-1536x864.jpg 1536w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-survey-380x214.jpg 380w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-survey-800x450.jpg 800w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-survey-1160x653.jpg 1160w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-survey-150x84.jpg 150w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-survey.jpg 1920w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p>Technical discovery fills in what people forget to mention. None of these sources is complete by itself, but together they usually surface more than firm leadership expects.<\/p>\n\n\n\n<figure class=\"wp-block-table\"><table class=\"has-fixed-layout\"><thead><tr><th>Source<\/th><th>What it catches<\/th><\/tr><\/thead><tbody><tr><td>A short staff survey<\/td><td>Tools people are already using but never thought to mention<\/td><\/tr><tr><td>Expense reports<\/td><td>Paid AI subscriptions nobody requisitioned through the firm<\/td><\/tr><tr><td>Browser extension reports<\/td><td>AI extensions installed on managed devices<\/td><\/tr><tr><td>Network traffic<\/td><td>Connections to known AI domains, where that visibility already exists<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p>Instead of guessing, the goal is a real inventory where the survey matters most. A technical scan tells you what happened on a managed device. It says nothing about the personal ChatGPT account someone\u2019s using from their phone.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"what-does-an-approved-tool-list-actually-look-like\"><strong>What does an approved-tool list actually look like?<\/strong><\/h2>\n\n\n\n<p><a href=\"https:\/\/www.forbes.com\/councils\/forbestechcouncil\/2026\/04\/09\/why-small-accounting-firms-face-more-ai-risk-than-large-ones\/\" target=\"_blank\" rel=\"nofollow noopener\"><strong>Verito CEO Jatin Narang laid out his own standard for vetting an AI tool in Forbes<\/strong><\/a><strong>: three questions, and none of them are answered by a sales page.<\/strong><\/p>\n\n\n\n<figure class=\"wp-block-table\"><table class=\"has-fixed-layout\"><thead><tr><th>Question<\/th><th>Why it matters<\/th><\/tr><\/thead><tbody><tr><td>Where does the data live?<\/td><td>U.S.-based servers keep the data inside a jurisdiction your existing compliance obligations already cover<\/td><\/tr><tr><td>Is the vendor SOC 2 compliant?<\/td><td>An independent audit, not a marketing claim, of how the vendor actually handles data<\/td><\/tr><tr><td>What do the retention and deletion terms say?<\/td><td>Determines whether client data sits inside the vendor\u2019s systems, or a training set, after the task is done<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p>Put the answer in writing. It doesn\u2019t need to be long, but it needs to name which tools are approved, for which tasks, and with which data, so \u201cprobably fine\u201d stops being the standard anyone\u2019s working from. Two more items belong on the same list: update engagement letters to address AI use where it touches client work, and ask the firm\u2019s E&amp;O carrier directly whether AI-related errors are covered, since most firm owners haven\u2019t asked and the answer isn\u2019t always what they\u2019d assume.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"where-policy-needs-technical-backing\"><strong>Where policy needs technical backing<\/strong><\/h2>\n\n\n\n<p><strong>A written policy works right up until someone\u2019s under deadline pressure and the sanctioned tool is slower than the one they already know. That\u2019s where the same category of control that blocks a phishing lure, network monitoring and content filtering, does double duty flagging or blocking traffic to AI tools that aren\u2019t on the approved list.<\/strong><\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"576\" src=\"https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-filtering-1024x576.jpg\" alt=\"Network control blocking unapproved AI traffic cinematic visual | Verito\" class=\"wp-image-8139\" srcset=\"https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-filtering-1024x576.jpg 1024w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-filtering-300x169.jpg 300w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-filtering-768x432.jpg 768w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-filtering-1536x864.jpg 1536w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-filtering-380x214.jpg 380w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-filtering-800x450.jpg 800w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-filtering-1160x653.jpg 1160w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-filtering-150x84.jpg 150w, https:\/\/verito.com\/blog\/wp-content\/uploads\/2026\/09\/shadow-ai-client-data-chatgpt-filtering.jpg 1920w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p>That written policy doesn\u2019t have to start from nothing, and it doesn\u2019t have to be built alone. If your firm is working out where to start, that\u2019s a conversation worth having directly.<\/p>\n\n\n\n<p><a href=\"https:\/\/verito.com\/contact\/\" target=\"_blank\" rel=\"dofollow\"><strong>Talk to us about building your firm\u2019s AI policy \u2192<\/strong><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"It\u2019s likely that someone at your firm has already pasted a client\u2019s information into ChatGPT, Claude, or Gemini&hellip;\n","protected":false},"author":12,"featured_media":8136,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1],"tags":[],"class_list":{"0":"post-8117","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-knowledge-base"},"acf":[],"_links":{"self":[{"href":"https:\/\/verito.com\/blog\/wp-json\/wp\/v2\/posts\/8117","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/verito.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/verito.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/verito.com\/blog\/wp-json\/wp\/v2\/users\/12"}],"replies":[{"embeddable":true,"href":"https:\/\/verito.com\/blog\/wp-json\/wp\/v2\/comments?post=8117"}],"version-history":[{"count":3,"href":"https:\/\/verito.com\/blog\/wp-json\/wp\/v2\/posts\/8117\/revisions"}],"predecessor-version":[{"id":8188,"href":"https:\/\/verito.com\/blog\/wp-json\/wp\/v2\/posts\/8117\/revisions\/8188"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/verito.com\/blog\/wp-json\/wp\/v2\/media\/8136"}],"wp:attachment":[{"href":"https:\/\/verito.com\/blog\/wp-json\/wp\/v2\/media?parent=8117"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/verito.com\/blog\/wp-json\/wp\/v2\/categories?post=8117"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/verito.com\/blog\/wp-json\/wp\/v2\/tags?post=8117"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}