If I were working for DEEPDREAM and I wanted to make sure that FB/T/YT/etc., were able to flag sensitive content automatically.. one thing I would do is scan 4 and 8 chan for all photo's, and mark down their names. And if anyone Tweets something without changing the filename, it will then zero in on it.. do a quick image checksum on it.. and flag it as sensitive (and also put you on a LIST because you are probably using the chan's as a source).

I have no proof they're doing this. But if I were building a system to do this, the Chan's would be obvious places to monitor in this way.

How to fight back?

Can Codemonkey or whoever controls this, automatically randomize image names before they get downloaded?

>>554566

BigData is all about efficiency. A filename would set off flags that set in motion a secondary measure to TinyEye the fuck out of the image just uploaded.

If Twitter were reverse image searching the millions of pictures uploaded every second, that would likely take a lot of resources.