Expired85% confidenceFactTime unknown
Prompt injection attacks could be used to induce ChatGPT connected to bank data to leak sensitive information such as users' account balances and transaction records.
1
Sources
85%
Confidence
Medium-term (~90 days)
Relevance
6/1/2026
First Seen
Valid until: 8/30/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedSensitive user information accumulated in noreply mailboxes can support identity theft, account takeover, and even financial fraud.68% similarVerified间接提示注入可将恶意指令藏入外部数据源（如网页、文档），待代理读取时自动触发攻击68% similarUnverified将登录后的ChatGPT页面内容复制粘贴到第三方网址会暴露会话令牌，攻击者可静默刷新access_token实现账号持续控制，即使修改密码在令牌被服务器撤销前依然有效63% similarUnverifiedUnder data protection regulations such as GDPR, sending personal data to the wrong email address may constitute a data breach incident.62% similarUnverified投资人在Twitter/X等平台公开背书已形成一套成熟的信号传递机制，可降低后续投资者的尽调信息不对称成本62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/22025API
curl https://kongchang.com/api/v1/knowledge/claims/22025MCP
get_claim(id=22025)