Imagine a user asks an LLM a question about a document. An attacker then gains an interactive prompt on the user's chat session, enabling the attacker to instruct the  AI assistant to take actions on behalf of the victim.

Join an in-depth roundtable session on a novel vulnerability class dubbed "Remote Prompt Execution" where, similar to remote code execution, an attacker is able to "run" arbitrary prompts on the user's chat session. Rubrik Zero Labs researchers will detail a full Copilot chain demonstrating a scenario where uploading a document to Copilot can lead to a full takeover of the Copilot's chat session.

Findings from this original research pointing to a number of Microsoft CVEs will be unveiled publicly for the first time at Black Hat 2026, and this is your opportunity to connect directly with the researchers responsible for the discovery. 

Join us for a roundtable lunch for a deep brief into the research findings.