The Synopsys Cybersecurity Research Center (CyRC) has exposed a data poisoning vulnerability in the EmbedAI application. EmbedAI allows users to interact with documents by utilizing the capabilities of large language models (LLMs).
This vulnerability could result in an application becoming compromised, leading to unauthorized entries or data poisoning attacks. These attacks are enabled by a cross-site request forgery (CSRF) vulnerability created by the absence of a secure session management implementation and weak cross-origin resource-sharing policies.
An attacker can direct a user to a malicious webpage that exploits a CSRF vulnerability within the EmbedAI application. By leveraging this CSRF vulnerability, the attacker can deceive the user into inadvertently uploading and integrating incorrect data into the application’s language model.
Exploitation of this vulnerability could affect the immediate functioning of the model and can have long-lasting effects on its credibility and the security of the systems that rely on it. This can manifest in various ways, including the spread of misinformation, introduction of biases, degradation of performance, and potential for denial-of-service attacks.
The CyRC reached out to the developers but has not received a response within the 90-day timeline dictated by our responsible disclosure policy. The CyRC recommends removing the applications from networks immediately.
This vulnerability was discovered by Mohammed Alshehri, a security researcher at Synopsys.
FIRST.Org, Inc (FIRST) is a non-profit organization based out of US that owns and manages CVSS. It is not required to be a member of FIRST to utilize or implement CVSS but FIRST does require any individual or organization give appropriate attribution while using CVSS. FIRST also states that any individual or organization that publishes scores follow the guideline so that anyone can understand how the score was calculated.