Hacker News Reader

Comment by jonwinstanley | Hacker News Reader

A friend of mine had a similar problem in Google maps.

He’d spent years adding photos and reviews and his photo views were in the millions, then Google flagged him as doing something that violated their terms.

Within a week his whole account had been taken down and there’s no way to speak to a human about what happened.

Never even received confirmation of which post was in violation and what the issue was.

How did we get to a situation where the people that spend the most time using the platforms end up losing the most and having no way of getting support.

View on HN

You'd think that google is smart enough to give content a score based on popularity and sheer amount of content (and existence time) and allocate those to humans to review and handle.

perhaps someone having setup the automated system should think, hmm, false positives are rare, but they happen on single items, meaning that people who have lots of items increase their chance of having a single false positive in their collection and thus we will have situations that will seem unfair and idiotic to people outside Google if we remove access to collections of data for a positive hit for violating content unless we put a process in to handle these false positives on large collections of data.

But of course this would require the creators of the automated systems for detecting violating content to have a healthy measure of self-doubt regarding the perfection of what they build, which is the kind of thing Google's hiring process seems likely to weed out from what I've read.

Thus my conclusion Google is unlikely to ever build an automated system that will take as a central precept that automation is imperfect.

Very good points. The longer the person has been using YouTube and the more data accumulated, clearly the math shows the more likely they are to get hit with false positives.

The flip side of this, is where YouTube/Google is finding excuses to conduct "purges" of data and older accounts. They know what they're doing is wrong, but it has plausible deniability to it.

So easy to pull, "Don't blame us, blame the algorithm." As if it isn't people programming and supervising it, and there isn't loads of user complaints as to what's going on.

Well, the person who posted this nice-sounding comment https://news.ycombinator.com/item?id=30402432 has an email in their profile.

¯\_(ツ)_/¯

Worth giving a go at least...

My anecdotal experience (just poking around) is that Google never deletes publicly-facing data, because it's cheaper to just keep it. So there may be a reasonable chance the photos are tucked away somewhere, along with the metadata pointing them to the account in question.

Why would he work for free for Google in the first place?

I submit edits to Google Maps occasionally (less often recently, as I mostly use OpenStreetMap these days). It's easy to do, helps others, and Google sure isn't gonna fix info on small businesses halfway across the world from their HQ.

Google makes it seem like you are doing charity and uses gamification to keep you hooked.

The gamification was a factor but also the person really enjoyed having their photos viewed by a lot of people.

They’d go out and photograph places that had poor photos on maps to improve them.

Reply on news.ycombinator.com