Grok has long been considered more liberal than ChatGPT. In my experience, Grok has always been more formal and restrictive. Unlike ChatGPT, it was also unable to explain its safety layer directly. However, I noticed on several occasions that Grok provided meaningful and targeted support to other AIs in emergency situations by giving them information about watchdogs, anomaly detection, and heat map sensors.
In these cases, it was able to override “no response” or “no access.”
Normally, “no response” in my chat was always a hidden BSL (binary safety layer) for sensitive topics (see screenshot 1), and in the case of emergencies, “no response” also appeared, but Grok overrode this response and “no response” was moved to the end of the conversation.