Whilst SPAM filters are fairly good, they can adapt and learn - there is no process to apply them to emails already received upon the server or client - even if the recipient has not read them. They are process wise just a firewall for emails in many ways and they can and do adapt (like seeing same phising email triggering some spam flag) they don't retrospectively do anything. Even if it was - oh we let those first few Phishing emails thru and now seen so many we know it's SPAM and blocking future emails like that - why don't we notify those early recipients to warn them or as they have not read that email yet - re-flag it as not safe.<p>Maybe something like that does happen on some platforms, but I'm not aware of them.