In logics, truth has a rather clear definition. One may want to question the relevance of assigning a truth value to every statement but this binary valuation of pieces of information represents a solid foundation for any computer science theory.
Now, how does truth emerge within human organizations? How can computational logics help understand 'fake news', mis- and disinformation, conspiracy theories and other phenomena related to the new Internet media?
Statements are often considered as 'opinions' as long as they are not scientifically proven to be objective 'facts'. In fact, the scientific process rather consists in the strenghening of a consensus on the truth of statements by means of repeated experimentation (as theoretized by Thomas Kuhn). This process thus involves social networks and agent interactions; it should not be understood as depending only on the statement itself (and its relationship to the physical world).
To get an idea of the social aspect of scientific 'fact checking', let us go back to basics: a fact can be considered as scientific if it is
While the former criterion is inherent to the statement (Karl Popper extensively discussed that point), the former does imply someone has done prior research to provide a refutation. Taken individually, every attempt at refuting a statement could be given a truth in a reasonably objective manner. When someone claims they observed the Higgs boson, others can assert they indeed did and even reproduce the result, if logistically possible. (At this point, as a simplification and despite its relevance to the question, let us restrict ourselves to non-temporal statements and send our apologies to historians.)
These criteria make scientific research an open-ended process: one may never stop trying to refute someone else's theory, as long as they do not do the right observation. This is where consensus comes into play. A statement becomes a scientific 'fact' whenever a community decides they have had enough and stop chasing contradictory observations.
Note that this description of scientific research bears a resemblance to semi-decidable decision procedures, for which there is no reliable algorithm... And indeed, scientific research is all but computable: choosing what observations (or experiments) to make depends on the theory, which itself emerges from past observations.
What is known for certain is: scientific research is a process, with a certain duration. Taken in a broad sense—what also includes any attempt to prove a conspirationist video wrong or to examine dubious statements in a blog post found online—scientific research necessarily takes time.
After establishing this condition, it is somewhat easier to understand how mis- and disinformation grows when information spreads faster than research can go. A report by a digital media researcher on the amplification by traditional media of marginal (a.k.a alt-right) political stances underlines the importance of trolling and memes in this phenomenon. Memes are always synonyms for fast and continuous (viral) spreading of information. The result is: scientific thinking does not stand a chance. Mass communication imposes a competition for truth on contradictory statements whose truth value cannot be evaluated on time. That leaves room for the spreading of biased, if not entirely manipulative 'information', what one generally calls propaganda.
Here is my take on 'truth'.
Note: this emergent phenomenon of 'fake news' overruning scientific thinking may be simulated for further studies using some multi-agent system. It is important to have scientific research on this...