And I find it hard to believe that there isn't any malware or phishing content on any of the big content provider's properties.
Realistically, offering an anonymous demo is probably a bad idea, no matter how many conversions it gives you.
No, but IF you allow people to anonymously upload malware to your site, Google-quality malware detection filters will absolutely do what they were designed to do and detect the malware on your site.
I just don't understand the people insisting on arguing that somehow this was a false positive. Google was right! This site was hosting malware! It's true it wasn't "intentionally" hosting malware, and that it was designed for benign purposes.
Which is to say, you are demanding that Google forgive this site automatically based on intent and not evidence. And sadly, while Google may have Google-quality malware detection filters, they haven't yet cracked the nut of thought crime detection.
I am genuinely curious how you would prevent your site from being blocked like this? Sure in this case it was a demo, but what if it was an actual image hosting service that required login? What's to stop a bad actor from creating an account and uploading malicious content. Maybe you even have some filter that does image recognition on the images and tries to detect if it is phishing. But unless your filter is able to catch the exact same content that google's malware detector uses, there's still a chance that you'll miss something that google finds, and starts blocking your site.
But even interpreting you narrowly: How much warning do you think Google should be expected to give before flagging sites with known malicious content? Would you apply that same logic to the sites you visit as a user?
I just don't see how the principle you want is going to work in concert with a world with rampant malware. Most of us very much want the trigger happy filers, because it keeps the problem manageable at the cost of some inconvenience and increased vigilance on the part of the content providers, which is IMHO exactly where it should be.
About half of them are taken down by the time I get to them. The other half, I've had to report myself.
I realize, however, that these are not malware.
(Funnily enough, I think all of the attempts I was able to view were trying to phish O365 creds.)