Unverified50% confidenceEventExact time
Attackers are attempting to manipulate large language models such as ChatGPT and Gemini into recommending, in their responses, URLs and contact information that lead to scam centers.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/24/2026
First Seen
Valid until: 12/23/2026
Sources
Related Entities
Related Claims
UnverifiedAttackers can induce agents to execute a sequence of seemingly normal operations—such as querying user data, calling formatting tools, and triggering email sending—to form a data exfiltration chain.67% similarVerifiedAI tools specifically customized for cybercrime, such as WormGPT and FraudGPT, have appeared on dark web forums and Telegram groups.66% similarUnverified2023年安全研究员Riley Goodside与Johann Rehberger等人演示了通过网页内容、邮件正文劫持GPT-4插件和AutoGPT的完整攻击链65% similarUnverifiedAttackers can use device search engines such as Shodan to locate S7 PLCs directly exposed to the internet.65% similarVerifiedPrompt injection attacks induce AI agents to execute malicious operations by embedding text disguised as system instructions in code comments, documentation, web content, or even file names.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/947206API
curl https://kongchang.com/api/v1/knowledge/claims/947206MCP
get_claim(id=947206)