[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fdgPIhZUccOCzczJVALnrJgIQO9TXJn4QyeU-WKe0Pfk":3},{"article":4,"iocs":48,"watch_terms":56,"solutions":57,"lesson":149},{"id":5,"title":6,"slug":7,"summary":8,"ai_summary":9,"brief":10,"url":11,"image_url":12,"published_at":13,"ingested_at":14,"relevance_score":15,"entities":16,"category_id":27,"category":28,"article_tags":32},"d9f321cd-3811-4a20-b527-f676eb616d59","ASCII smuggling crosses over from AI prompt injection to phishing evasion","ascii-smuggling-crosses-over-from-ai-prompt-injection-to-phishing-evasion-a60156","Invisible Unicode characters popularized for hiding instructions from AI models are now being used to obfuscate words before email filters parse them. The post ASCII smuggling crosses over from AI prompt injection to phishing evasion appeared first on Microsoft Security Blog.","Microsoft researchers have observed a high-volume phishing campaign utilizing invisible Unicode tag characters, a technique previously popularized in AI prompt injection research known as ASCII Smuggling. Attackers are now employing this method to split financial lure words, preventing email filters from parsing them. This demonstrates how AI-era evasion techniques are being adapted for traditional phishing campaigns, with Microsoft Defender for Office 365 seeing a sharp increase in related detections.","Attackers use AI prompt injection technique to evade email filters in phishing campaigns.","https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fsecurity\u002Fblog\u002F2026\u002F09\u002F03\u002Fascii-smuggling-crosses-over-from-ai-prompt-injection-to-phishing-evasion\u002F","https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fsecurity\u002Fblog\u002Fwp-content\u002Fuploads\u002F2026\u002F03\u002FMS_Actional-Insights_Adversarial-AI.jpg","2026-09-03T16:00:00+00:00","2026-09-03T18:00:19.525807+00:00",8,[17,20,23,25],{"name":18,"type":19},"Microsoft Defender for Office 365","product",{"name":21,"type":22},"AI prompt injection","technology",{"name":24,"type":22},"ASCII Smuggling",{"name":26,"type":22},"Unicode","80544778-fabb-4dcd-aa35-17492e5dcf4f",{"id":27,"icon":29,"name":30,"slug":31},null,"Vulnerabilities","vulnerabilities",[33,38,43],{"category":34},{"id":35,"icon":29,"name":36,"slug":37},"839da5c1-3c34-47e2-9499-f7201640e3ac","AI Security","ai-security",{"category":39},{"id":40,"icon":29,"name":41,"slug":42},"89f78b1c-3503-45a1-9fc7-e23d2ce1c6d5","Malware","malware",{"category":44},{"id":45,"icon":29,"name":46,"slug":47},"e7b231c8-5f79-4465-8d38-1ef13aea5a14","Threat Intelligence","threat-intelligence",[49,53],{"type":50,"value":51,"context":52},"mitre_attack","T1566.002","Phishing: Spearphishing Attachment",{"type":50,"value":54,"context":55},"T1027","Obfuscated Files or Information",[18],{"controls":58,"vendors":142},[59,87,116],{"slug":47,"name":60,"problem":61,"products":62},"Threat intelligence","Attack campaigns reach your sector, suppliers or software before your team hears about them.",[63,71,79],{"vendor_slug":64,"vendor_name":65,"vendor_logo":66,"product_slug":67,"product_name":68,"url":69,"description":70},"google","Google","\u002Fvendor-logos\u002Fgoogle.png","google-threat-intelligence","Google Threat Intelligence","https:\u002F\u002Fcloud.google.com\u002Fsecurity\u002Fproducts\u002Fthreat-intelligence","Threat intelligence from Mandiant's frontline investigations, Google's threat analysts, open sources and Google's own visibility, in one service.",{"vendor_slug":72,"vendor_name":73,"vendor_logo":74,"product_slug":75,"product_name":76,"url":77,"description":78},"crowdstrike","CrowdStrike","\u002Fvendor-logos\u002Fcrowdstrike.png","falcon-adversary-intelligence","Falcon Adversary Intelligence","https:\u002F\u002Fwww.crowdstrike.com\u002Fen-us\u002Fplatform\u002Fthreat-intelligence\u002F","Threat intelligence on the adversaries CrowdStrike tracks, built into the Falcon platform to detect, disrupt and defend against them.",{"vendor_slug":80,"vendor_name":81,"vendor_logo":82,"product_slug":83,"product_name":84,"url":85,"description":86},"dataminr","Dataminr","\u002Fvendor-logos\u002Fdataminr.png","agentic-threat-intelligence-platform","Agentic Threat Intelligence Platform","https:\u002F\u002Fwww.dataminr.com\u002Fproducts\u002Fcyber-defense\u002Fagentic-threat-intelligence-platform\u002F","Threat intelligence platform, formerly ThreatConnect TI Ops, that collects, scores and structures intelligence and pushes it into security tools.",{"slug":88,"name":89,"problem":90,"products":91},"email-security","Email security","Phishing, malware and invoice fraud are filtered out before they reach inboxes.",[92,100,108],{"vendor_slug":93,"vendor_name":94,"vendor_logo":95,"product_slug":96,"product_name":97,"url":98,"description":99},"cloudflare","Cloudflare","\u002Fvendor-logos\u002Fcloudflare.png","cloudflare-email-security","Cloudflare Email Security","https:\u002F\u002Fwww.cloudflare.com\u002Fsase\u002Fproducts\u002Femail-security\u002F","Email filtering for phishing and business email compromise.",{"vendor_slug":101,"vendor_name":102,"vendor_logo":103,"product_slug":104,"product_name":105,"url":106,"description":107},"mimecast","Mimecast","\u002Fvendor-logos\u002Fmimecast.png","mimecast-email-security","Mimecast Email Security","https:\u002F\u002Fwww.mimecast.com\u002Fproducts\u002Femail-security\u002F","Email filtering, impersonation protection and continuity.",{"vendor_slug":109,"vendor_name":110,"vendor_logo":111,"product_slug":112,"product_name":113,"url":114,"description":115},"abnormal","Abnormal AI","\u002Fvendor-logos\u002Fabnormal.png","abnormal-email-security","Abnormal Inbound Email Security","https:\u002F\u002Fabnormal.ai\u002F","API-based email security for Microsoft 365 and Google Workspace using behavioural models.",{"slug":117,"name":118,"problem":119,"products":120},"security-awareness-training","Security awareness training and phishing simulation","People recognise and report attacks instead of falling for them.",[121,129,136],{"vendor_slug":122,"vendor_name":123,"vendor_logo":124,"product_slug":125,"product_name":126,"url":127,"description":128},"huntress","Huntress","\u002Fvendor-logos\u002Fhuntress.png","managed-sat","Huntress Managed Security Awareness Training","https:\u002F\u002Fwww.huntress.com\u002Fplatform\u002Fsecurity-awareness-training","Security awareness training and phishing simulation, managed for you.",{"vendor_slug":130,"vendor_name":131,"vendor_logo":132,"product_slug":130,"product_name":133,"url":134,"description":135},"sosafe","SoSafe","\u002Fvendor-logos\u002Fsosafe.png","SoSafe Awareness Platform","https:\u002F\u002Fsosafe-awareness.com\u002F","Behaviour-based awareness training and phishing simulation in many languages.",{"vendor_slug":137,"vendor_name":138,"vendor_logo":139,"product_slug":137,"product_name":138,"url":140,"description":141},"hoxhunt","Hoxhunt","\u002Fvendor-logos\u002Fhoxhunt.png","https:\u002F\u002Fhoxhunt.com\u002F","Adaptive phishing simulations and micro-training, with a focus on reporting.",[143],{"slug":144,"name":145,"logo_path":146,"roles":147},"microsoft","Microsoft","\u002Fvendor-logos\u002Fmicrosoft.png",[148],"reported_by",{"slug":150,"title":151,"body":152,"framework_refs":153,"framework_notes":163},"invisible-unicode-characters-enable-phishing-filter-evasion-via-ascii-smuggling","Invisible Unicode Characters Enable Phishing Filter Evasion via ASCII Smuggling","Attackers are exploiting invisible Unicode tag characters to split financial lure keywords in phishing emails, bypassing traditional email content filters that cannot parse these hidden characters. This technique, originally developed in AI prompt injection research, demonstrates how offensive methods rapidly migrate from one threat domain to another — in this case from AI jailbreaking to conventional phishing campaigns. The core failure is that email security tooling was not designed to account for non-rendering Unicode characters as a text obfuscation vector. This matters because organizations relying solely on keyword-based filtering are now blind to a growing category of evasion techniques, leaving employees exposed to convincing financial-themed lures. The sharp increase in detections by Microsoft Defender for Office 365 signals this is an active, scaling threat requiring immediate defensive adaptation.",[154,155,156,157,158,159,160,161,162],"CIS Control 9: Email and Web Browser Protections","CIS Control 17: Security Awareness and Skills Training","NIST SP 800-61 Rev 2: Incident Handling for Phishing","NIST SP 800-177: Email Authentication Recommendations","NIST CSF DE.CM-7: Monitoring for Unauthorized Activity","MITRE ATT&CK T1027: Obfuscated Files or Information","MITRE ATT&CK T1566.001: Spearphishing Attachment","GDPR Article 32: Security of Processing (if PII is at risk via phishing)","ITIL: Continual Service Improvement — adapting controls to emerging threats",{}]