What you are describing here, is mostly the operation of confirmation bias, not evidence. To try to understand things, we build rival abstract explanatory models. At its best, evidence, is something that is independent of human opinion that identifies which of the explanatory models is truest; that is why the scientific approach is founded on the elimination of subjective human input. It's not possible to eliminate humans from the loop completely though; it's always human minds that conceive the models and human minds interpret results. Human minds are infested with legacy beliefs and inherited biases and we cannot help but see the world the the lens of the portfolio of biases that we inherit. Maybe when we succeed in building conscious machines, they will tell us what is really true and what is not. Or maybe they too will become infected with our biases
Your confirmation bias is incredible! You seriously think truth is only objective truth....and subjectivity is a hindrance.
All objective observation and analysis (even by a robot) has to be limited by its sensory inputs and the software that is loaded in it. Even a robot has to be programmed to analyse data. The programming is its limitation....besides its limited sensory devices. I am not talking about bias...but about natural limitations.
'Truth' has to exist beyond these limitations.
If someone introduces data regarding a phenomenon that lies beyond the observation of the robot....the robot will not be able to analyse the data free of its limitation. It will still produce the same result or treat the data as redundant because that's what it is programmed to do.
Robots will be more limited than humans because they cannot change their programs internally as humans can through their will ..... with just one small trigger. We humans are more than just sensory machines.