Mark Zuckerberg has big plans for regulating 'hate speech' through AI on Facebook

Mark Zuckerberg has big plans for regulating "hate speech" on Facebook.  And he's stepping up artificial intelligence (A.I.) operations to silence those he deems hate-speakers, to keep the community "safe."  He forecasts that it will be a done deal in five to ten years.  Asked how he evaluates hate speech and where he draws the line on it at Facebook, by Republican Sen. John Thune of South Dakota during testimony in the Senate yesterday, Zuckerberg replied:

YES, MR. CHAIRMAN. I'LL SPEAK TO HATE SPEECH, THEN I'LL TALK ABOUT ENFORCING OUR content POLICIES MORE BROADLY. SO ACTUALLY, MAYBE IF YOU ARE OKAY WITH IT, I'LL GO IN THE OTHER ORDER. SO FROM THE BEGINNING OF THE COMPANY, IN 2004, I STARTED IN MY DORM ROOM, IT WAS ME AND MY ROOMMATE, WE DIDN'T HAVE AI TECHNOLOGY THAT COULD LOOK AT CONTENT that PEOPLE WERE SHARING. SO WE BASICALLY HAD TO ENFORCE OUR CONTENT POLICIES REACTIVELY. PEOPLE COULD SHARE WHAT THEY WANTED, AND THEN IF SOMEONE IN THE COMMUNITY FOUND IT TO BE OFFENSIVE or against our POLICIES THEY WOULD FLAG IT FOR US AND WE WOULD LOOK AT IT REACTIVELY. NOW INCREASINGLY, WE ARE DEVELOPING AI TOOLS THAT can identify certain classes of bad activity proactively and FLAG IT for our team AT FACEBOOK. BY THE eND OF THIS YEAR by the way WE'LL HAVE MORE THAN 20,000 PEOPLE WORKING ON SECURITY AND CONTENT REVIEW WORKING ACROSS ALL THESE THINGS. SO WHEN CONTENT GETS FLAGGED TO US WE HAVE THOSE PEOPLE LOOK AT AND IF IT VIOLATES OUR POLICIES THEN WE TAKE IT down. SOME PROBLEMS LEND THEMSELVES MORE EASILY TO AI SOLUTIONS THAN OTHERS. SO HATE SPEECH IS ONE OF THE HARDEST BECAUSE DETERMINING IF SOMETHING IS HATE SPEECH IS VERY LINGUISTICALLY NUANCED. YOU NEED TO UNDERSTAND WHAT IS A SLUR AND WHETHER SOMETHING IS HATEFUL. NOT JUST IN ENGLISH, BUT MAJORITY OF PEOPLE ON FACEBOOK USE IT IN LANGUAGEs that are DIFFERENT ACROSS THE WORLD. CONTRAST that FOR EXAMPLE WITH AN AREA LIKE FINDING TERRORIST PROPAGANDA WHICH WE'VE actually BEEN VERY SUCCESSFUL AT DEPLOYING AI TOOLS ON ALREADY. TODAY AS WE SIT HERE 99% OF THE ISIS AND AL QAEDA content that WE TAKE DOWN on Facebook, our AI systems FLAGS BEFORE ANY HUMAN SEES IT. SO THAT'S SUCCESS IN TERMS OF ROLLING OUT AI TOOLS THAT CAN PROACTIVELY POLICE AND ENFORCE SAFETY ACROSS THE COMMUNITY. HATE SPEECH, I'M OPTIMISTIC THAT OVER A FIVE TO TEN YEAR PERIOD, WE'LL HAVE AI TOOLS THAT CAN GET INTO SOME OF THE NUANCES, the LINGUISTIC NUISANCES, OF DIFFERENT TYPES OF CONTENT TO BE MORE ACCURATE IN FLAGGING THINGS FOR OUR SYSTEMS but TODAY WE ARE just NOT THERE on that. SO A LOT OF THIS IS still REACTIVE. PEOPLE FLAG IT TO US. WE HAVE PEOPLE LOOK AT IT. WE HAVE POLICIES to try to make as not-subjective as possible. BUT UNTIL WE Get it MORE AUTOMATED THERE IS ERROR RATE I'M NOT HAPPY WITH. >> THANK YOU.

I corrected the transcript and highlighted the money quotes.  Here is a C-SPAN video clip of the same:

Thune wasn't able to follow up on the key thing, which is finding out how Zuckerberg defines "hate speech," and the only information Zuck provided was "slurs" and speech that "is hateful."  Does he mean that anyone who criticizes President Obama or Hillary Clinton is a practitioner of hate speech?  And as for slurs, if he really meant it, shouldn't this mean that quite a few Democrats would get shut down as avid practitioners?  Hard to believe.  They don't even get shut down for assassination talk.

Apparently, Zuckerberg enforces his hate-speech policies based on "if someone found it offensive," meaning that a lot of snowflakes are in business.  Snowflakes find everything offensive.  Will he go by the snowflake standard?  We all suspect what the answer will be.

He's vague about "bad activity," and that is worth noting, too, given that groups such as the Southern Christian Leadership Conference consider pretty much anything conservative "hate speech" and brands it as such.  SCLC is a vaunted institution of the establishment, oh, so respectable, and yet it does that because the establishment is extremist and hates Republicans with a passion.  Is Zuckerberg going to be listening to people like that?  Don't bet on him not listening to them.

He's also hiring a big police force to police speech, and while many of them will be policing foreign-language posts, it should have been asked how many of them would be policing American posts written in English.  The number contrast would be interesting.  One guy for the Telugu desk and 15,000 for the Republicans?  Lucky us.

Zuck does note that policing hate speech is one of the hardest issues for his company, without quite saying why.  Yet it contrasts easily with Facebook's A.I. detecting terrorist activity, probably because all you need are keywords for that, and you are targeting a specific entity, not a "nuanced" person you don't know who lives a normal life in the community and just doesn't like Eric Holder or Hillary Clinton.  I'll give him some pointers here as to why it's hard: it's because censoring Republicans brings bad publicity, does it not?  And we know that's what he has in mind.

He gets more Orwellian when he talks about enforcing "safety" in speech, and that immediately calls to mind the Facebook shadowbanning or "deplatforming" of Diamond & Silk, two pro-Trump Republican women whose content was censored out of their subscribers' news feeds because, as Facebook put it, they were "unsafe to the community."  Someone should have asked him point-blank if he considers the two black women who love Trump "terrorists."

He then pays lip service to making his so-called hate speech enforcement as non-subjective as possible.  Coming from a leftist such as Zuckerberg who considers left-wing shibboleths on everything from global warming to Obamacare to illegal aliens having more rights than U.S. citizens as absolute truths, this doesn't inspire confidence.

The bottom line here is that Zuckerberg is planning to deploy artificial intelligence tools, programmed by their human programmers, into censorship of Republican speech, in the name of left-wing truisms about what truth is.  If that's not a direct campaign contribution to Democrats, and evidence of an edited published platform that needs to be regulated for content, as newspapers are, then what is?  Regulate them.  Zuckerberg flunked his Senate testimony and deserves nothing less.

Mark Zuckerberg has big plans for regulating "hate speech" on Facebook.  And he's stepping up artificial intelligence (A.I.) operations to silence those he deems hate-speakers, to keep the community "safe."  He forecasts that it will be a done deal in five to ten years.  Asked how he evaluates hate speech and where he draws the line on it at Facebook, by Republican Sen. John Thune of South Dakota during testimony in the Senate yesterday, Zuckerberg replied:

YES, MR. CHAIRMAN. I'LL SPEAK TO HATE SPEECH, THEN I'LL TALK ABOUT ENFORCING OUR content POLICIES MORE BROADLY. SO ACTUALLY, MAYBE IF YOU ARE OKAY WITH IT, I'LL GO IN THE OTHER ORDER. SO FROM THE BEGINNING OF THE COMPANY, IN 2004, I STARTED IN MY DORM ROOM, IT WAS ME AND MY ROOMMATE, WE DIDN'T HAVE AI TECHNOLOGY THAT COULD LOOK AT CONTENT that PEOPLE WERE SHARING. SO WE BASICALLY HAD TO ENFORCE OUR CONTENT POLICIES REACTIVELY. PEOPLE COULD SHARE WHAT THEY WANTED, AND THEN IF SOMEONE IN THE COMMUNITY FOUND IT TO BE OFFENSIVE or against our POLICIES THEY WOULD FLAG IT FOR US AND WE WOULD LOOK AT IT REACTIVELY. NOW INCREASINGLY, WE ARE DEVELOPING AI TOOLS THAT can identify certain classes of bad activity proactively and FLAG IT for our team AT FACEBOOK. BY THE eND OF THIS YEAR by the way WE'LL HAVE MORE THAN 20,000 PEOPLE WORKING ON SECURITY AND CONTENT REVIEW WORKING ACROSS ALL THESE THINGS. SO WHEN CONTENT GETS FLAGGED TO US WE HAVE THOSE PEOPLE LOOK AT AND IF IT VIOLATES OUR POLICIES THEN WE TAKE IT down. SOME PROBLEMS LEND THEMSELVES MORE EASILY TO AI SOLUTIONS THAN OTHERS. SO HATE SPEECH IS ONE OF THE HARDEST BECAUSE DETERMINING IF SOMETHING IS HATE SPEECH IS VERY LINGUISTICALLY NUANCED. YOU NEED TO UNDERSTAND WHAT IS A SLUR AND WHETHER SOMETHING IS HATEFUL. NOT JUST IN ENGLISH, BUT MAJORITY OF PEOPLE ON FACEBOOK USE IT IN LANGUAGEs that are DIFFERENT ACROSS THE WORLD. CONTRAST that FOR EXAMPLE WITH AN AREA LIKE FINDING TERRORIST PROPAGANDA WHICH WE'VE actually BEEN VERY SUCCESSFUL AT DEPLOYING AI TOOLS ON ALREADY. TODAY AS WE SIT HERE 99% OF THE ISIS AND AL QAEDA content that WE TAKE DOWN on Facebook, our AI systems FLAGS BEFORE ANY HUMAN SEES IT. SO THAT'S SUCCESS IN TERMS OF ROLLING OUT AI TOOLS THAT CAN PROACTIVELY POLICE AND ENFORCE SAFETY ACROSS THE COMMUNITY. HATE SPEECH, I'M OPTIMISTIC THAT OVER A FIVE TO TEN YEAR PERIOD, WE'LL HAVE AI TOOLS THAT CAN GET INTO SOME OF THE NUANCES, the LINGUISTIC NUISANCES, OF DIFFERENT TYPES OF CONTENT TO BE MORE ACCURATE IN FLAGGING THINGS FOR OUR SYSTEMS but TODAY WE ARE just NOT THERE on that. SO A LOT OF THIS IS still REACTIVE. PEOPLE FLAG IT TO US. WE HAVE PEOPLE LOOK AT IT. WE HAVE POLICIES to try to make as not-subjective as possible. BUT UNTIL WE Get it MORE AUTOMATED THERE IS ERROR RATE I'M NOT HAPPY WITH. >> THANK YOU.

I corrected the transcript and highlighted the money quotes.  Here is a C-SPAN video clip of the same:

Thune wasn't able to follow up on the key thing, which is finding out how Zuckerberg defines "hate speech," and the only information Zuck provided was "slurs" and speech that "is hateful."  Does he mean that anyone who criticizes President Obama or Hillary Clinton is a practitioner of hate speech?  And as for slurs, if he really meant it, shouldn't this mean that quite a few Democrats would get shut down as avid practitioners?  Hard to believe.  They don't even get shut down for assassination talk.

Apparently, Zuckerberg enforces his hate-speech policies based on "if someone found it offensive," meaning that a lot of snowflakes are in business.  Snowflakes find everything offensive.  Will he go by the snowflake standard?  We all suspect what the answer will be.

He's vague about "bad activity," and that is worth noting, too, given that groups such as the Southern Christian Leadership Conference consider pretty much anything conservative "hate speech" and brands it as such.  SCLC is a vaunted institution of the establishment, oh, so respectable, and yet it does that because the establishment is extremist and hates Republicans with a passion.  Is Zuckerberg going to be listening to people like that?  Don't bet on him not listening to them.

He's also hiring a big police force to police speech, and while many of them will be policing foreign-language posts, it should have been asked how many of them would be policing American posts written in English.  The number contrast would be interesting.  One guy for the Telugu desk and 15,000 for the Republicans?  Lucky us.

Zuck does note that policing hate speech is one of the hardest issues for his company, without quite saying why.  Yet it contrasts easily with Facebook's A.I. detecting terrorist activity, probably because all you need are keywords for that, and you are targeting a specific entity, not a "nuanced" person you don't know who lives a normal life in the community and just doesn't like Eric Holder or Hillary Clinton.  I'll give him some pointers here as to why it's hard: it's because censoring Republicans brings bad publicity, does it not?  And we know that's what he has in mind.

He gets more Orwellian when he talks about enforcing "safety" in speech, and that immediately calls to mind the Facebook shadowbanning or "deplatforming" of Diamond & Silk, two pro-Trump Republican women whose content was censored out of their subscribers' news feeds because, as Facebook put it, they were "unsafe to the community."  Someone should have asked him point-blank if he considers the two black women who love Trump "terrorists."

He then pays lip service to making his so-called hate speech enforcement as non-subjective as possible.  Coming from a leftist such as Zuckerberg who considers left-wing shibboleths on everything from global warming to Obamacare to illegal aliens having more rights than U.S. citizens as absolute truths, this doesn't inspire confidence.

The bottom line here is that Zuckerberg is planning to deploy artificial intelligence tools, programmed by their human programmers, into censorship of Republican speech, in the name of left-wing truisms about what truth is.  If that's not a direct campaign contribution to Democrats, and evidence of an edited published platform that needs to be regulated for content, as newspapers are, then what is?  Regulate them.  Zuckerberg flunked his Senate testimony and deserves nothing less.