This might not even be a problem that needs to be solved client side.
Relays already have their own spam protection mechanisms.
Discussion
Spam has a broad definition. Relays could handle obvious spam attacks (e.g. bots sending the same link as a reply to each thread), but in the end it is the user, hence the client, who should choose what to filter.
But the possibility of automating the creation of new profiles exists anyway, so I agree this is not an argument against simplifying onboarding.