I think a much better way to do this would be to send a confirmation email first. After the user accepts the confirmation email then it continues with the submission. This would prevent potentially evil stuff from happening.

Please don't get me wrong, I didn't write this, and don't endorse it, but I find it interesting.

My suspicion echos yours, particularity based on the name (A play on http://en.wikipedia.org/wiki/Jailbait perhaps?)

That said, I think it's an interesting project, and could legitimately be quite useful in the right circumstances.

I'm a SysAdmin- I've had to train and test SPAM filters quite a few times in setting up new systems. (Zimbra/etc)

Creating a throw-away address on the new system, and having this blast it with relevant spam would be very helpful to me.

I do agree they ought to have a confirm-your-email link, though, to prevent abuse.

You can't train spam filters with 100% spam. Moreover you should train with the spam your users actually receive.

No, but you can look at what got through and figure out why it got through, then adjust settings accordingly. You might find you need to give more weight to specific blacklists or something.

mailbait is not intended to signup for spam. The email returned is solicited from legitimate services; including listservs, newsletters, and anything else users dig up to add to the system. This makes it a good fit for some tests, as inbound email is not necessarily from illegitimate sources with negative intentions.

Surely this breaks some TOSs right? Marketers are going to lose revenue by spamming people that don't want it, and they're essentially a spam service. This is pathetic.