Hammering a controversial term like 'white genocide' into a system prompt with specific orders creates a fixation effect in the AI. Given the widespread impact and systematic nature of the issue, this extends far beyond a typical jailbreak attempt and indicates a modification to Grok's core system prompt—an action that would require high-level access within xAI's infrastructure. "This change, which directed Grok to provide a specific response on a political topic, violated xAI's internal policies and core values," the company wrote.
Author: Jose Antonio Lanz
Published at: 2025-05-17 19:00:02
Still want to read the full version? Full article
