Papers looking at whether or not this is desirable tend not to come to that conclusion, e.g. https://www.tandfonline.com/doi/abs/10.1080/08913811.2013.84... https://www.tandfonline.com/doi/abs/10.1080/10584609.2019.16... (2020 meta-analysis) https://www.tandfonline.com/doi/abs/10.1080/15205436.2020.17...
I personally hold a belief that a fact-checking publisher has the same biases and profit-motives as any news publisher (or, as motive, political influence of the wealthy).
I don't find it useful to distinguish fact-checking as special and call for specific research on this, as if existing academic research can always be invalidated by a simple re-branding.
I don't know how well this will age, but for now, I think any company claiming to be able to automate fact-checking should be met with a high degree of skepticism.
Just for context, what I imagine in my head when I think of a "fact check" is something like this, basically a specialized or somewhat standardized / formulaic form of journalism:
- https://www.snopes.com/fact-check/england-banned-fox-news/
- https://www.politifact.com/factchecks/2021/feb/18/joe-biden/...
- https://www.politifact.com/factchecks/2021/jan/05/donald-tru...
I deliberately picked a couple of less obvious fact-checks than "is the earth round", because I wanted to highlight how even in politics, on questions where there is high potential for bias, there are cut and dry things which can be fact-checked, or at the very least 'contextualized'.
I'm sure none of this is perfect, but I am not super optimistic that automated content would be of this level of quality.