A Design Flaw in Share Links Exposes Users’ Personal Conversations With Meta’s Muse
Key points
- Researchers from security firm Claroc say Meta's Muse AI assistant exposed sensitive conversations shared through web links. They found 6,800 share links and opened a few hundred, revealing items including a loan application, a bank fraud alert and a staff vacation calendar.
- The links were easy to find because the conversation title appeared in the web address, which public DNS records captured. Meta said it has paused the sharing feature while it reviews the issue and considers changes.
- Earlier, a separate vulnerability reported by security expert Patrick Wardle had also been downplayed by Meta.
Get local alerts + the full feed in the AU Now app
Alerts, Shorts and your saved stories — free on Google Play.
Disclaimer
This is an AI-generated summary based on reporting by Wired AI. It may not reflect every detail of the original story exactly. Tap “Continue reading” above for the full, original article.
Follow this story in the AU Now app
Alerts for your suburb, Shorts and saved stories — free on Google Play.
Comments
No comments yet — start the conversation.