It's often thought that epistemic responsibility is (largely) a matter of bringing one's beliefs "in line" with the evidence. Specifically, it is a matter of bringing one's beliefs in line with the evidence that bears on the truth of those beliefs. Now, according to Williamson, for any condition C such that one must not believe p unless condition C obtains, one ought not believe p unless one has adequate evidence that C does in fact obtain. To do otherwise would be negligent at best. (See his remarks about the truth norm for assertion in Knowledge and its Limits. I think it's around pp. 246, but that's from memory.) So, quick argument that knowledge is not the norm of belief. To believe p without adequate evidence for the second-order belief that this belief constitutes knowledge is not epistemically irresponsible. But, it would be were knowledge the norm of belief. So, it's not knowledge that's the norm of belief. Truth is the norm of belief.
I think we can bolster this point as follows. According to the knowledge account, this talk of "adequate evidence" is to be understood in terms of evidence adequate for knowledge. If you think that it is permissible to believe p if p is known and deny KK principles, we have a problem. If you deny that knowledge of p's truth requires the availability of evidence that puts you in a position to know that you are in a position to know p (the denial of KK), then you either have to say that permissible belief doesn't require "adequate evidence" in the relevant sense or allow for the possibility of permissibly believing in an epistemically irresponsible way.
Similar problems arise for Huemer in his defense of the knowledge norm. His view appears to be that you shouldn't believe something that does not satisfy a comprehensive epistemic endorsement and that a belief can only be properly endorsed in this way if that belief constitutes knowledge. So, suppose we say of someone that while they know p they are not in a position to know that they know this. Here's the problem. According to Huemer, I'm permitted to hold the second-order belief that my belief about p can be comprehensively endorsed only if the second-order belief constitutes knowledge. If it does not, I oughtn't believe the first-order belief can be comprehensively endorsed and thus oughtn't hold that belief. So, either some KK principle is true or knowledge of p's truth is insufficient for permissibly believing p.
I don't see that parallel problems arise if we say that the fundamental norm of belief is the truth norm.