Has anyone been working on automatic generation of QA review questions
weighted by their real-world outcome preference implications?
I recently noticed a controversy about when rebar for reinforced
concrete was invented. I had looked the fact up earlier out of idle
curiosity, but I realize that many insurance plans are struggling with
the end-of-life claims on reinforced concrete foundations, so there is
a potential financial incentive to mislead people about the date that
rebar was invented: it would be easier to terminate insurance
contracts and total more buildings, which has substantial dollar
figure financial implications.
I am not suggesting the process of ranking questions by their dollar
value implications could be automated without human support, but has
anyone explored the extent to which automation can assist? Updating is
a core, if not the most core as of last year, editing activity
necessary to maintain quality. Wikidata can help, but QA review needs
to be independent and relatively anonymous.