A threat actor has claimed to have exploited an Insecure Direct Object Reference (IDOR) vulnerability in AI research platform Silvi AI, allegedly exposing more than 16,000 user records. The claims have not been independently verified, although the actor published technical details describing the purported attack.
According to the forum post, the vulnerability allowed authenticated users to retrieve other users’ profile information by iterating through sequential account IDs. The actor claims to have extracted data belonging to 16,483 users after creating a standard account through the platform’s public registration endpoint.
IDOR vulnerability allegedly enabled user enumeration
Silvi AI is a Denmark-based platform that helps researchers conduct AI-assisted literature reviews and academic research.
The threat actor alleges the issue stemmed from an API endpoint that returned user information based solely on a supplied numeric ID, without validating whether the authenticated user was authorized to access that record. According to the post, the actor created a new account, obtained a JSON Web Token (JWT), and then queried user IDs sequentially until approximately 16,483 records had been retrieved.
If accurate, the described behavior is consistent with an IDOR vulnerability, a common access control flaw that can expose data when object-level authorization checks are missing.
Claimed data includes user identities
The actor claims the exported records contain email addresses, first and last names, internal user IDs, organization identifiers, subscription status, and profile image locations where available.
BreachNews reviewed the sample published alongside the claim, which appears consistent with the fields described in the post. However, the publication of sample records alone is insufficient to independently verify the scope or authenticity of the alleged breach.
Latest AI platform targeted
The alleged compromise is the latest in a series of claims involving AI-focused services. Earlier this week, BreachNews reported on another alleged breach attributed to the same threat actor involving Darsa AI, where source code, credentials, and internal files were reportedly exposed. Read our coverage of the Darsa AI incident here.
At the time of publication, Silvi AI had not issued any public statement regarding the alleged incident. BreachNews will update this article if the company confirms the breach or additional evidence becomes available.
