It's a reasonable point but this would probably be a losing battle; at this point terms like "classical security", "classical adversaries", etc. are common in the literature.

To me what's worse is "zk" used to describe applications of verifiable computation with no secrets involved, but that seems like a losing battle also.