> you have no way to tell if the info is correct or not, and the prisoner has no incentive to give you the correct information
(In the following, I'm trying to just discuss the technical aspects of this. When I ask if or suggest that something would be effective, it is not meant to imply that if torture can be made effective then it is OK to do it. The ineffectiveness of torture as currently practiced is just one argument against torture, and so if it can be made effective that still leaves the other anti-torture arguments intact).
Wouldn't this largely depend on what type of information you are trying to get and how much you already know? If you already have information that I know that you are not supposed to know, and I don't know you already have it, you can ask about that while interrogating me in addition to asking about whatever it is you are really trying to get. That should give you some feedback on whether or not I'm giving correct info.
Wouldn't it also depend on how many other people you are interrogating over the same subjects? If you are questioning me and several other people about a particular thing, and I make up something on the fly to get you to stop torturing me, my made up story might not be the same as the made up stories of other people, whereas the stories of the people who tell the truth will agree. Of course, we could all have been trained to expect to be interrogated over this, and all have consistent prepared lies to give. Your counter to that would be to try to capture and interrogate lower level people or people who were less directly involved so that they are less likely to have prepared stories to give.
I can believe that torture as currently practiced is almost always ineffective, but I suspect that this may be due to them relying on psychologists to figure out how to do it rather than bringing in engineers, scientists, and mathematicians and treating it as some kind of noisy and unreliable channel problem.