Ask him to explain how he thinks he can exploit it. Post that explanation. It might be worthwhile for him to read the reference docs for the class first, rather than relying entirely on your informal description.
I've never heard of any exploits that solely involved NSNotificationCenter. That is, an exploit where NSNotificationCenter and only that class was the primary cause of an exploited vulnerability. If there is such an intrinsic vulnerability, I'd be interested in hearing what it is.
NSNotificationCenter is per-process. It's not the same as NSDistributedNotificationCenter, which goes across processes, but is still constrained to notifications on the same machine. And NSDistributedNotificationCenter isn't the same as Distributed Objects. In any case, if an attacker can execute arbitrary code on a target machine, there are a lot simpler avenues than attacking NSNotificationCenter. If the attacker can't execute arbitrary code, then that rules out a lot of potential attacks, regardless of what they might attempt.
It's always possible to write a program that does something foolhardy when it receives a notification. That would be a vulnerability in the recipient's action, though, not in NSNotificationCenter itself. Example: a notification observer deletes any pathname sent to it in the userInfo dictionary, without first ascertaining that the sender is authorized.