“Safe” the platonic ideal is an impossibility. Any cryptographic solution depends ultimately on handshake agreements between fallible human executives and/or fallible human registries, and there’s no known alternative to that today. Is RPKI “safe”, relative to not RPKI? Yes, obviously, it is. Is it reasonable to interpret “safe” as ‘no further improvement is required’? Never: this is the Internet; one could expect the domain to be repurposed to cover more than RPKI someday. Yes, short-sighted leaders may use “RPKI is safe” as justification to withhold investment forward past it; but that outcome is certain regardless of how they justify it.