Your best friend, Sally, has just been acquitted of murder. A remote acquaintance of yours, Ravi, has a justified true belief that Sally is innocent. Ravi is mistrustful of the legal system and the fact that Sally was acquitted carries no weight. But what does carry weight for him is that he believes his friend Patricia had seen Sally at a McDonald's at exactly the time the crime was being committed (and, no, he never told the authorities). Patricia, however, did not see Sally at McDonald's—she saw someone else who looked like her from the distance (a distance normally sufficient for distinguishing people, but this lookalike was pretty close). On the other hand, you know that Helga really did see Sally at a Burger King at that time, and indeed saw her closely enough that Ravi's report is not a defeater for Helga. So, Ravi does not know that Sally is innocent, while both you and Helga do.
If knowledge has a value over and beyond that of justified true belief, then Ravi is missing out on something of value that Helga has. This would imply that you have a reason out of charity for Ravi to tell him that Patricia did not see Sally, but Helga did. Moreover, you have two reasons out of charity for Ravi to tell him this. One reason is generated simply by the value of replacing a false belief (that Patricia saw Sally) with a true one (that Helga saw Sally). This reason is pretty weak, since the question of who saw Sally is intrinsically pretty unimportant, and while all truth has value, unimportant truths have but little value. The second reason is there if and only if knowledge has a value over beyond justified true belief: this is the value of Ravi's knowing Sally to be innocent, rather than merely having a justified true belief about it.
Oddly enough, I find myself having the intuition that I have a moderately strong reason to ensure that Ravi knows my best friend Sally to be innocent. What is odd about it is that this intuition conflicts with my theoretical views on which knowledge, as compared to mere justified true belief, has only instrumental value (potential to generate more in the way of true beliefs, etc.). When I initially set out the case, I thought I would have the intuition that I only have a weak reason to tell Ravi that he's wrong about Patricia seeing Sally.
Maybe, though, I can reconcile with my intuitions as follows. Maybe the moderately strong reason I have is generated merely by the reason in charity to replace Ravi's false belief about Patricia seeing Sally with a true belief about Helga seeing Sally. It would seem like this belief is fairly inconsequential in itself, and so the reason generated this way would be weak. But maybe the importance of a belief depends in part on the importance of the things derived from it. And since Sally's innocence is important, Ravi's false belief is important. If so, then the case doesn't undercut my theoretical views.
And here's a further intuition: It is very important that Ravi believe Sally to be innocent; the value of his knowing Sally to be innocent, even if greater, is only a little greater.