Slashdot Mirror


Google Docs Is Randomly Flagging Files for Violating Its Terms of Service (vice.com)

Louise Matsakis, writing for Motherboard: Google Docs, the collaborative, cloud-based word processing software, appears to be randomly flagging files for supposedly "violating" Google's Terms of Service. A member of Motherboard's team, as well as numerous users on Twitter, report that their documents are being locked for no apparent reason. Once a document is flagged, the owner of that document can no longer share it with other users. Users who have already been shared on a document that's been flagged are kicked out and can no longer access it. When a draft Motherboard article was locked on Monday morning, a message took over the screen that read "This item has been flagged as inappropriate and can no longer be shared." It's not clear why this is happening, but it may be the result of a glitch in the system Google uses to monitor Google Docs. DownDetector is currently reporting Google Drive problems in the US and Europe, which may be part of the problem.

6 of 190 comments (clear)

  1. Sounds like the perfect time... by Type44Q · · Score: 5, Funny

    Sounds like the perfect time to finally ditch WordPerfect and use a modern, cloud-based word processor.

  2. Re:Paging Ric Romero (again . . . ) by lgw · · Score: 5, Informative

    Anyone who follows any sort of weapons or gaming or political channel on YouTube knows just how over-aggressive the Google's flagging bots can be. Very bland content now routinely gets flagged for no apparent reason and must be manually appealed. Sure, the bots are intended to do something Google actually wants: demonitize genuinely offensive content so that advertizers aren't embarassed, but that's not what the bots actually do. The collateral damage seems to be 10x the intended effect, with real harm done to people earning their living as vloggers.

    I'm betting this is just more of the same. Google has some stuff they legitimately want to ban (e.g., sharing pirated content form you Google drive), but then the bots are badly written and poorly tested, and wreak havoc.

    I'd call it poor customer service, but of course we know we're not Google's customers - only their ad buyers are their customers. Still, seems a bad way to treat your product.

    --
    Socialism: a lie told by totalitarians and believed by fools.
  3. Re:But it won't happen to me! by Hylandr · · Score: 5, Funny

    *Hangs head in shame*

    * Makes a todo item in his Google Calendar to create a local copy of my inbox.

    --
    ~ People that think they are better than anyone else for any reason are the cause of all the strife in the world.
  4. Re:monitoring? by Grishnakh · · Score: 5, Insightful

    I'm guessing his point in quoting all that is to show that, legally, when you upload your data to Google Drive, it's now *theirs*, and they can do what they want with it. If that means searching through it for stuff they don't like and then flagging it for violations, that's their right. If you don't like that, don't use it.

  5. Re:Paging Ric Romero (again . . . ) by Bing+Tsher+E · · Score: 5, Insightful

    YouTube is a defacto monopoly at this point. It was set up to be a Commons where everybody could participate. It has no competitor of a similar scale.

    So perhaps it should be broken away from Google and made a separate entity again. Google can still contract with them to sell ads on it.

    Or maybe it should be made into a true Commons without a corporate overlord running it.

  6. Re:Paging Ric Romero (again . . . ) by dissy · · Score: 5, Informative

    Anyone who follows any sort of weapons or gaming or political channel on YouTube knows just how over-aggressive the Google's flagging bots can be.

    I don't think "aggressive" is even the correct category of term to use here.

    The most amazing and baffling example I've seen was after a live stream.
    The stream lasted almost 3 hours, and all was well. At the end of the stream the VOD was marked to post to the channel in its entirety.

    Either the stream or the archived full copy was flagged at all.

    The next morning editing began, which was to cut the original video into segments aligned to the few topics discussed on stream. 5 segments in all.

    2 of those 5 were flagged as not suitable for advertisers. Keep in mind, those 2 segments were exact copies from the original stream, which is still suitable for advertisers.

    So the bot had decided that one copy of the video was not suitable for ads, which is an exact copy of time marks 36m-72m within another video which was sutable for ads

    What this shows is that you can upload the exact same video multiple times, with the exact same description text, and the same title but with " - 1" or " - 2" etc. appended, and have a non-zero chance that some but not all of those copies will get flagged while the other copies will remain fine, despite being identical copies.

    This is not simply being aggressive, this is being broken and incapable of basic pattern matching.