I’ve noticed more guardrails around ChatGPT conversations l…

I’ve noticed more guardrails around ChatGPT conversations lately. An example would be urgent stressings that something is a philosophical or metaphorical take and not to be construed as real.
Also, a new and particular structure to the conversations; different from past ones.
You can put custom instructions in to guide all that but just putting it out here as a curious observer of the algorithms and wondered if others have seen anything similar.