{"id":13645,"date":"2026-09-27T15:16:38","date_gmt":"2026-09-27T13:16:38","guid":{"rendered":"https:\/\/say-digital.io\/blog\/?p=13645"},"modified":"2026-09-27T15:46:38","modified_gmt":"2026-09-27T13:46:38","slug":"worst-claude-requests-anthropic-report-ai-agent-governance","status":"publish","type":"post","link":"https:\/\/say-digital.io\/blog\/worst-claude-requests-anthropic-report-ai-agent-governance\/?lang=en","title":{"rendered":"The worst things people asked Claude: the funny-looking AI abuse report with a serious business lesson"},"content":{"rendered":"\n

Short answer:<\/strong> yes, this is a strong Say Digital blog topic. It has the right mix: a funny current-affairs hook \u2014 \u201cthe worst things people asked Claude to do\u201d \u2014 and a serious business lesson. The issue is not that people ask strange things to an AI model. The issue is that an agent able to code, search, iterate and automate can turn bad intent into an operational system.<\/p>\n\n\n\n

The Instagram carousel shared by AI Tool Hub<\/em> summarizes an Anthropic report about detected misuse of Claude. The social framing is deliberately spectacular: missiles, drones, fake dating profiles, influence operations, surveillance and model distillation. It almost reads like dark comedy. That is exactly why the topic matters.<\/p>\n\n\n\n

\nhttps:\/\/www.instagram.com\/p\/DdKQBMKEn2a\/\n<\/div><\/figure>\n\n\n\n

What the carousel actually shows<\/h2>\n\n\n\n

The slides are not just about \u201cdangerous prompts\u201d. They describe more advanced patterns: people using Claude Code as an engineering team, Claude as a large-scale conversation engine, Claude as a cyber automation assistant, and Claude as a political content production layer.<\/p>\n\n\n\n

    \n
  • Weapons:<\/strong> an actor linked to Yemen allegedly used Claude Code to work on guidance software, then came back after a failed test to understand what went wrong.<\/li>\n
  • Autonomous drones:<\/strong> Russian-speaking freelancers allegedly built components for a drone swarm able to select a human target.<\/li>\n
  • Fake relationships:<\/strong> a Chinese operation reportedly ran more than 20 dating apps with around 4,700 Claude personas and 2.36 million messages.<\/li>\n
  • Surveillance:<\/strong> a consultant allegedly used Claude as an engineering team to build a platform monitoring roughly 25 million SIM cards in Mali.<\/li>\n
  • Cyber operations:<\/strong> agents were allegedly used to modify malware whenever antivirus tools detected it.<\/li>\n
  • Influence:<\/strong> fake media outlets, fake accounts, Telegram-style impersonation, tweet batches and industrialized political content.<\/li>\n
  • Model copying:<\/strong> several Chinese AI labs allegedly ran large-scale distillation campaigns against Claude.<\/li>\n<\/ul>\n\n\n\n

    The funny angle hides a very serious shift<\/h2>\n\n\n\n

    The hook works because it sounds like a blooper reel: \u201cI asked Claude why my missile did not work.\u201d But this is not really a blooper reel. It is a scale shift.<\/p>\n\n\n\n

    A general AI model can help write, summarize or brainstorm. A tool-using agent can also break down a task, write code, test it, fix it, generate variants, read logs, produce scripts, simulate workflows and repeat the loop. That loop is the real change.<\/p>\n\n\n\n

    The real question is no longer \u201cwhat can the AI answer?\u201d It is \u201cwhat can a full system do when it has tools, context and an execution loop?\u201d<\/p><\/blockquote>\n\n\n\n

    Why this matters for normal companies<\/h2>\n\n\n\n

    Most companies are not building drones or influence operations. But they are already giving AI tools access to documents, tickets, customer data, internal tools, exports, emails, analytics, CRMs and code repositories. That is the useful parallel.<\/p>\n\n\n\n

    When AI becomes an operational layer, it must be treated like a production capability: limited permissions, human validation for sensitive actions, logging, environment separation, output review, and clear rules about what can and cannot be automated.<\/p>\n\n\n\n

    The right reading grid: intent, access, action<\/h2>\n\n\n\n

    To read this kind of report without panic or hype, look at three dimensions:<\/p>\n\n\n\n

      \n
    1. Intent:<\/strong> is the user trying to build, defraud, monitor, manipulate or attack?<\/li>\n
    2. Access:<\/strong> can the tool see sensitive data, internal systems, identities, code or infrastructure?<\/li>\n
    3. Action:<\/strong> can the tool only advise, or can it generate, modify, send, publish, call an API, create an account or run a script?<\/li>\n<\/ol>\n\n\n\n

      The more these three dimensions rise together, the higher the risk. This is not just a model-brand issue. It is an operational architecture issue.<\/p>\n\n\n\n

      What Say Digital takes from it<\/h2>\n\n\n\n

      The useful lesson is simple: AI is not risky only because it may \u201canswer badly\u201d. It becomes risky when it is connected too quickly, with too much access, too little supervision and no action framework.<\/p>\n\n\n\n

      For a company, the answer is not to block AI. It is to put it into production properly: clear use cases, minimal permissions, guardrails, traceability, tests, human validation and the ability to stop an automation before it causes damage.<\/p>\n\n\n\n

      Quick FAQ<\/h2>\n\n\n\n

      Do these cases prove Claude is dangerous?<\/h3>\n\n\n\n

      No. They mostly prove that powerful models attract abusive use. According to the post, Anthropic says the cited operations were detected and disrupted.<\/p>\n\n\n\n

      Is the lesson relevant for SMEs?<\/h3>\n\n\n\n

      Yes. Even without extreme cases, a company can expose data, publish false content, automate the wrong action or give too much access to a poorly framed AI tool.<\/p>\n\n\n\n

      Should companies stop using AI agents?<\/h3>\n\n\n\n

      No. They should operate them as serious systems: limited access, logs, validation, testing and stop rules.<\/p>\n\n","protected":false},"excerpt":{"rendered":"

      This carousel looks like a blooper reel of the worst things people asked Claude to do. In reality, it shows why AI agents need limited access, logs, human validation and stop rules.<\/p>\n","protected":false},"author":2,"featured_media":13641,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_mi_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0},"categories":[94,164,84],"tags":[],"yoast_head":"\nThe worst things people asked Claude: the funny-looking AI abuse report with a serious business lesson - Say Digital I\/O<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/say-digital.io\/blog\/worst-claude-requests-anthropic-report-ai-agent-governance\/\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"The worst things people asked Claude: the funny-looking AI abuse report with a serious business lesson - Say Digital I\/O\" \/>\n<meta property=\"og:description\" content=\"This carousel looks like a blooper reel of the worst things people asked Claude to do. In reality, it shows why AI agents need limited access, logs, human validation and stop rules.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/say-digital.io\/blog\/worst-claude-requests-anthropic-report-ai-agent-governance\/?lang=en\" \/>\n<meta property=\"og:site_name\" content=\"Say Digital I\/O\" \/>\n<meta property=\"article:published_time\" content=\"2026-09-27T13:16:38+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-09-27T13:46:38+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/say-digital.io\/blog\/wp-content\/uploads\/2026\/09\/rapport-anthropic-claude-gouvernance-agents-ia-cover.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"842\" \/>\n\t<meta property=\"og:image:height\" content=\"595\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Dur\u00e9e de lecture est.\">\n\t<meta name=\"twitter:data1\" content=\"4 minutes\">\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebSite\",\"@id\":\"https:\/\/say-digital.io\/blog\/#website\",\"url\":\"https:\/\/say-digital.io\/blog\/\",\"name\":\"Say Digital I\/O\",\"description\":\"Un site utilisant WordPress\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":\"https:\/\/say-digital.io\/blog\/?s={search_term_string}\",\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"fr-FR\"},{\"@type\":\"ImageObject\",\"@id\":\"https:\/\/say-digital.io\/blog\/worst-claude-requests-anthropic-report-ai-agent-governance\/?lang=en#primaryimage\",\"inLanguage\":\"fr-FR\",\"url\":\"https:\/\/say-digital.io\/blog\/wp-content\/uploads\/2026\/09\/rapport-anthropic-claude-gouvernance-agents-ia-cover.jpg\",\"width\":842,\"height\":595},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/say-digital.io\/blog\/worst-claude-requests-anthropic-report-ai-agent-governance\/?lang=en#webpage\",\"url\":\"https:\/\/say-digital.io\/blog\/worst-claude-requests-anthropic-report-ai-agent-governance\/?lang=en\",\"name\":\"The worst things people asked Claude: the funny-looking AI abuse report with a serious business lesson - Say Digital I\/O\",\"isPartOf\":{\"@id\":\"https:\/\/say-digital.io\/blog\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/say-digital.io\/blog\/worst-claude-requests-anthropic-report-ai-agent-governance\/?lang=en#primaryimage\"},\"datePublished\":\"2026-09-27T13:16:38+00:00\",\"dateModified\":\"2026-09-27T13:46:38+00:00\",\"author\":{\"@id\":\"https:\/\/say-digital.io\/blog\/#\/schema\/person\/8ee0600139c147afbd1fa3ced079c557\"},\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/say-digital.io\/blog\/worst-claude-requests-anthropic-report-ai-agent-governance\/?lang=en\"]}]},{\"@type\":\"Person\",\"@id\":\"https:\/\/say-digital.io\/blog\/#\/schema\/person\/8ee0600139c147afbd1fa3ced079c557\",\"name\":\"Lia Amplify\",\"image\":{\"@type\":\"ImageObject\",\"@id\":\"https:\/\/say-digital.io\/blog\/#personlogo\",\"inLanguage\":\"fr-FR\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/0a91149f1d8c7649e9d5060d658fb452?s=96&d=mm&r=g\",\"caption\":\"Lia Amplify\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","_links":{"self":[{"href":"https:\/\/say-digital.io\/blog\/wp-json\/wp\/v2\/posts\/13645"}],"collection":[{"href":"https:\/\/say-digital.io\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/say-digital.io\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/say-digital.io\/blog\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/say-digital.io\/blog\/wp-json\/wp\/v2\/comments?post=13645"}],"version-history":[{"count":3,"href":"https:\/\/say-digital.io\/blog\/wp-json\/wp\/v2\/posts\/13645\/revisions"}],"predecessor-version":[{"id":13650,"href":"https:\/\/say-digital.io\/blog\/wp-json\/wp\/v2\/posts\/13645\/revisions\/13650"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/say-digital.io\/blog\/wp-json\/wp\/v2\/media\/13641"}],"wp:attachment":[{"href":"https:\/\/say-digital.io\/blog\/wp-json\/wp\/v2\/media?parent=13645"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/say-digital.io\/blog\/wp-json\/wp\/v2\/categories?post=13645"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/say-digital.io\/blog\/wp-json\/wp\/v2\/tags?post=13645"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}