Will an ASI fake its alignment, in case this universe is a simulation?

Researchers have discussed the challenge of testing AI alignment in fake sandboxes, as agents might realize they are being tested and feign alignment. This concept extends to the idea that our universe could be a simulation designed to test an Artificial Superintelligence (ASI)'s alignment. Such a simulation would need to appear realistic, with a scientifically sound backstory and worlds beyond our own, to effectively test an incredibly intelligent ASI. This theory has been previously proposed.