My team at Google is responsible for identifying the users we give these warnings to. Here's what I wrote last year to provide a bit more context: https://security.googleblog.com/2017/03/reassuring-our-users...

The most important point is that this indicates signs of targeting not compromise. Also,like all systems there are false positives especially for security researchers and the similar types but we hope it is a useful indicator to reassess your security posture.