3 ms·

More evidence that automated systems like this should not be used.

• jagged-chisel 6y ago
  I don't think it's the automated detection that's a problem. It's the "you cannot object, you have no recourse, you cannot reply, you cannot show your license, and when we decide this has happened enough times, you cannot use our platform" that's the problem.
  • kodt 6y ago
    Yes, an automated system that flags content but is human reviewed would be more reasonable. But the human review part slows things down and you end up losing much of the advantage of the automated system as you would be throttled by human review or have to spend much more on manpower.
• Spivak 6y ago
  Why? Their purpose is to stop platforms from getting sued for copyright violations. They are remarkably successful to that end. Their goal isn’t to allow all content that wouldn’t get them sued but to catch all content that would. False positives get you angry messages on Twitter but false negatives get you multimillion dollar lawsuits.