As long as they're careful to never lose video that may become the subject of media attention, I suppose it's possible. That seems like a game that's guaranteed to end badly, though, because it's not possible to know what will and will not become big news. Some things are obviously big (e.g. deaths) but lesser issues may not blow up until some subsequent sequence of events.
It seems like a really risky business strategy for Taser... and if any journalist ever caught wind of the "soft points", or any whistleblower decided to out them, it'd generate a firestorm.
What you describe is very feasible in heavily manual processes, but automated systems operate in the way they're designed, without variation, and deliberate holes leave traces. With manual processes, the worst case is that the agency has to find a scapegoat, some low-level employee who was responsible for doing something and didn't. With automated systems, it's much harder to argue that the problem wasn't deliberate. It's easy enough the first time "It's a bug!", but it doesn't take long before people want to know why the bug hasn't been fixed.