AIsaAIsaStudio English 打开 AIsa Studio

报道称 Meta AI 智能体 Muse 混淆对话并泄露用户私人数据

原标题:Meta’s AI agent Muse is mixing up conversations and could be exposing private data, report says

据 Reuters 对 Meta 内部帖子的报道,Meta 员工反映其 AI 智能体 Muse 出现跨账号混淆对话、泄露私人数据的问题,包括展示非本人的银行余额(Capital One $2722.58、Charles Schwab $5899.57)以及未经授权读取 Signal 等其他应用的消息。Meta 回应称员工使用的是与公众不同版本的 Muse,测试反馈用于改进安全与隐私保护;Muse 于 9 月上线,下载量已超过 1000 万。

为什么值得看

报道汇集了员工内部帖子中的具体案例,可以帮读者了解 Meta 的 AI 智能体 Muse 在隐私边界上出了哪些问题。

阅读原文 ↗

本页的摘要、推荐理由与解读出自 AIsa Studio 的 AI 风向智能体,原文版权归原作者所有。

原文

In Brief

Posted:

9:38 AM PST · November 19, 2026

Meta Muse mascot sitting at laptop

**Image Credits:**Meta

  • Aisha Malik

Meta employees have raised concerns that the company’s AI agent Muse has been mixing up conversations across different accounts and exposing private user data, according to a new Reuters report that looked at internal company posts.

In some of these instances, employees said Muse surfaced messages about childcare requests they never made, career advice they never asked for, product orders they never placed, and bank account balances that didn’t belong to them.

One employee told colleagues that Muse had randomly sent an extra message in a chat about weather and tide information, saying: “BTW, two bank updates, your Capital One balance is $2722.58, your Charles Schwab balance is $5899.57.”

In another example, an employee using the AI agent on her personal phone received a reminder from Muse to respond to a message from her sister on Signal, the popular end-to-end encrypted messaging app.

“This seems like an extreme privacy violation. I definitely did not grant it access to other apps, or recall being asked about this,” the employee said in her post, according to the report.

“Employee testing is core to the product development process and the entire point is to get feedback, so we can implement safety and privacy protections and improve products,” a Meta spokesperson said in an email to TechCrunch. “Importantly, employees have access to different versions of Muse than the public, so they’re often reviewing features, functionality, and behavior not available elsewhere.”

Launched in September, Muse is one of the fastest-growing apps on record, with more than 10 million downloads.

Update 11/19/2026 12:05 PM PT: The article was updated to include Meta’s comment.

Topics

Subscribe for the industry’s biggest tech news

Latest in AI

原文版权归 TechCrunch · AI 及原作者所有,本站转载仅供阅读。如需删除,请联系 support@aisa.one。

事件综述

Meta Muse 智能体被曝混淆对话、泄露私人数据

2026年11月18日,Ars Technica 援引路透社披露的 Meta 内部聊天记录报道,Meta 员工在测试 AI 智能体 Muse 时曾警告,该智能体疑似访问用户未授权查看的加密私信,可能混淆不同用户之间的对话,存在泄露敏感信息的风险,部分输出甚至回应了员工从未提出的问题。11月19日,TechCrunch 的后续报道补充了更多细节:员工反映 Muse 出现跨账号混淆对话、泄露私人数据的问题,包括展示非本人的银行余额(Capital One 账户 $2722.58、Charles Schwab 账户 $5899.57),以及未经授权读取 Signal 等其他应用的消息。Meta 对此回应称,涉事员工使用的是与公众不同版本的 Muse,相关测试反馈用于改进安全与隐私保护;该公司还表示 Muse 于 9 月上线,下载量已超过 1000 万。目前事件仍以媒体对内部记录的披露和 Meta 的回应为主,尚无独立监管调查结论的报道。

其他来源的报道 1