Governments creating AI that can block “extremist” content because website businesses aren’t doing it fast or well enough. This is problematic for two reasons:
- “Extreme” is a point of view. When this is implemented points of view not in line with businesses and governments never see the light of day, and
- Governments and businesses become the judges of what is permissible, legal or illegal, not due process.
- The AI algorithms work very, very well, and will allow businesses and governments to pinpoint even as much as someone’s emotional state. I recommend Zeynep Tufeckci’s TED talk, https://www.whisperloudly.net/brilliant-ted-talk-worse-than-1984/)
We’ve had people spouting off on street corners for literally thousands of years and people just walk on by, some listen, some don’t. When a government car pulls up and those people are shoved into the back of the car and they disappear without due process, it’s obviously a human rights violation. This situation that’s rapidly becoming mainstream differs little.
So what if “extremist” videos are aired? People need to be educated and it should be discussed, this is cultural. We’re rapidly becoming a culture where citizens see what governments and businesses want them to see, not given the option to think for themselves. “Fake News” is a slippery slope – control the news people see and don’t allow posting of contrary opinion and your control is complete.
Lastly, it’s a small step forward to extend this to law: A short future from now, if you upload a video that the AI improperly flags as “extreme” or otherwise illegal, your house could get raided based on that information alone (It’s almost never wrong, after all), and now it’s your responsibility to defend yourself and get your stuff (and your life!) back. “Innocent until proven guilty” is bullshit nowadays; people get picked up and jailed all the time, and the impetus is on them to come up with bail and lawyers to prove otherwise.
It’s not that far a stretch, here are some supporting articles on that vein:
- https://www.alternet.org/civil-liberties/america-innocent-until-proven-guilty-not-most-people-who-are-stuck-jail
- http://www.slate.com/articles/technology/future_tense/2016/12/software_problems_are_leading_to_wrongful_arrests.html
- https://arstechnica.com/tech-policy/2016/12/court-software-glitches-result-in-erroneous-arrests-defense-lawyers-say/
To quote Ms. Bitters (Invader Zim) – “SILENCE! The machine has spoken!”