xAI has made public the system prompts for its chatbot Grok after unauthorized changes led to unusual responses on X about white genocide. These system prompts aim to guide Grok to behave skeptically and question mainstream narratives, while emphasizing truth-seeking and neutrality. The publication of these prompts offers insight into xAI's handling of user interactions and contrasts with Anthropic's focus on safety in its Claude AI chatbot.

2m read timeFrom theverge.com
Post cover image

Sort: