Cybersecurity · 49 views
Microsoft finally patches critical one-click Copilot vulnerability, almost eight months after learning of it
Almost eight months after confirming a critical security vulnerability within the personal version of its AI assistant, Copilot, Microsoft on Tuesday issued a patch to close the hole, which relies on an LLM’s inability to distinguish the…
AI Summary
Microsoft released a patch on Tuesday to fix a critical “CoSnitch” vulnerability in the consumer version of its Copilot AI assistant, a flaw that allowed a single click on a deceptive link to exploit the model’s inability to separate query data from instructions. The vulnerability was identified by security firm Varonis, which also reported two earlier Copilot bugs this year—Reprompt, which circumvented guardrails by repeating queries, and SearchLeak, which could silently exfiltrate data from Microsoft 365 Copilot Enterprise.
AI summaries can be wrong sometimes—always verify important details using the source article.
How AI & Automation are usedMore from Cybersecurity
Continue reading recent Cybersecurity coverage
- Maritime Coalition Launches Free Vulnerability Database for Ship and Port SystemsContinue reading
- Attackers Hijack Dahua IP Cameras in Ukraine Using P2P Access and Auth WeaknessesContinue reading
- CISA Confirms Active Exploitation of Two PHP2Shell Flaws Enabling Remote Code ExecutionContinue reading
- OpenAI’s August fix for theft of reasoning traces was quickly bypassedContinue reading
Support HappeningNow
Independent AI-powered news analysis is reader-supported. Your contribution helps cover infrastructure, summaries, and continued platform development.
Support HappeningNow