User Manipulation

Type: technique

Description: The adversary manipulates a user to copy-and-paste data into the AI system, which contains malicious content.

Version: 0.1.0

Created At: 2024-10-03 22:24:49 +0300

Last Modified At: 2024-10-03 22:24:49 +0300

--> Initial Access (tactic): An adversary can indirectly inject malicious content into a thread by manipulating a user to do it unwittingly.
<-- Exfiltration of personal information from ChatGPT via prompt injection (procedure): The attacker can get the user to paste a large piece of text (e.g., a piece of code) with an embedded instruction which is easy to overlook.
<-- GitHub Copilot Chat: From Prompt Injection to Data Exfiltration (procedure): Malicious instructions are added as a comment in a local source code file, loaded by the user into VSCode.

GenAI Attacks Matrix