About failing checks in LGTM

Though i am not a fan of using LGTM, but while browsing through the reports, i found out there are instances that actually makes sense. 
My question revolves around whether we should ask the respective code authors to do the same, in a way relaying to them the code quality issue, which can be used as learning and as a reference for new contributors, or do it all by ourselves.