Reference on how often the AFAIKs turn out to be true

Quite often authors make some claims with some "to the best of our knowledge" disclaimer. E.g. "To the best of our knowledge this is the first time this approach has been used to address this problem".

Is there any research/study/survey that tried to look at how often such claims are true?

I am mostly interested in the field of computer science > machine learning / NLP / data mining, in English-speaking venues.