In my experience as dungeon master, players are very accepting of "I don't know" as a response. But as soon as you "mislead" players, they start to think of that situation as a betrayal of trust.
So "villagers don't know" whats ahead is far better than "villager MISTAKENLY thinks there's a vampire ahead".
Do you really want to run games where the players spend 10+ rolls asking different villagers for randomly generated knowledge, because they know that some knowledge is tainted?
Do you, as the dungeon master, place more tainted knowledge into the pool, or do you place more correct knowledge in the pool?
If the pool of knowledge is 50% tainted, why bother even listening to villagers in the first place? Its clear they don't know what they're talking about. Save yourself an hour of pointless roleplaying / acting out NPCs / etc. etc. and just say "The villagers don't know".
------------
I mean, I've had players who don't like combat and prefer running around the world doing... not adventures or something. But they're in the exceptional minority in my experience. Most players want to know where the monster is, which direction to go to fight the monster, and then to save the town. Its cheesy, but its what works.
----------
I have successfully run a "whodunit" scenario. Where I explicitly lay out 5 NPCs, with their own incomplete view of some situation. And the players need to talk with the NPCs at specific times (and being seen with some NPCs conflicts with other meetings. Two NPCs are enemies and will refuse to cooperate if they see you conversing with their rival).
If the players fail to find the antagonist out of the pool, no big deal. Have a "one step backwards" adventure (ex: Culann's Hound situation or similar), and then take on the antagonist afterwards.
The story must go onward. Not necessarily always in the benefit of the players, but it takes a lot of effort to craft an actually fun, good, and engaging set of "unreliable NPCs" that people actually want to play with.
Hey, you do your campaigns as you wish! Just noting that what you call 'tainted information' is part of color and challenge. Use it how you like.
Or more importantly: is talking to NPCs in your video game more fun than top-tier visual novels (Ex: Phoenix Write: Ace Attorney, "When they Cry", or Danganronpa )?
As a player, I know what I'm getting into. I choose to play mindless hack-and-slash games because the mindless hack-and-slash is quite fun sometimes. I don't play Dynasty Warriors for in-depth stories or deep NPC character development.
In contrast, when I pickup a visual novel game like Danganronpa, I'm not really going to be wowed by difficult reflexes or hand-eye coordination.
Games exist as a drop of culture in a greater ecosystem. The ultimate truth is, your video game plotlines are never going to be as deeply engaging as a book's plot.
After all, a book has literally nothing else to go on aside from character development and interesting plot development. People play video games (or tabletop games) for different reasons.
It's personal preference if you lean more into action or social, but I find either end of the spectrum to be to limited to just one thing.
I've read plenty of dull books, and played many engagingly plotted games. But then those games are often criticized as having too much text.
Part of the joy of games is agreement on the parameters of what is and is not in bounds for the game. Knowing that boundary lets you focus on only the things within that while pleasantly mentally unloading all of the myriad messy things outside.
Even though each videogame is different, there is a surrounding videogame metagame culture that players tacitly assume defines what all videogames are allowed to do. When learning a new videogame, players assume the rules of that metagame still hold true. (This is one of the reasons videogames today can be so difficult for non-gamers to get into: they don't have this absorbed culture.)
One of the rules of that metagame is "NPCs do not mislead the player." (Other rules of the metagame are: "Your savefile cannot get into a state that makes it impossible to beat the game," and "It should not be required to die/fail several times in order to learn how to proceed.")
It's valid for a game to break that meta-rule, but it will be very unpleasantly surprising if they believe they are playing the stock videogame metagame as they learn this game's rules.
This is kind of unfortunate because there's a whole giant space of possible videogames that could be made with interesting rules if only you could set the expectation for players that this game steps outside of typical game culture meta rules. You occasionally see experimental games that do this like Lose/Lose, which deletes random files off your hard drive when you lose (!).
I played a game a while back that made a sort of Groundhog Day repeated dying a requirement to make progress. That was a central theme to the game, so it ensured the player understood the situation pretty quickly.
E.g. if one villager said the mirror is in the swamp, unlucky players might waste a lot of time in the swamp, which is bad. But saying it's an axe vs a sword vs a piece of armor vs a potion probably doesn't matter much - whatever it is, as long as it is what is required to defeat the dragon, it's fine.
Can work well for a puzzle though. Recall a few Zelda games having setups where one NPC was actually a spy for a rival group, and you needed to cross examine what other NPCs said to figure out who that was.