AI and the future of censorship
One of the reasons why the internet has become such a bastion of free speech is that there is such a massively huge volume of stuff on the internet that it’s impossible for human censors to censor everything (even though Facebook has been trying by hiring thousands of human censors).
The future of censorship is AI-based, and it’s very scary indeed. Once AI can be trained to detect “racist” writings, it can go through the billions of postings on Facebook, Twitter, etc., and delete everything “racist.” Google could even put this technology into Gmail, thus preventing you from sending or receiving any email that’s deemed “racist.”
Today we live in a bubble of free speech, but it’s a temporary bubble.
Lion of the Blogosphere 
I don’t feel that worried about this particular kind of censorship yet.
As long as you aren’t using the N-word it would be hard for an AI bot to really know what you are talking about.
An AI bot could either err on the side of free speech or err on the side of censorship. If the former, it would be easy to avoid by avoiding key triggering words. If the latter, they quickly have a problem with their users.
I think the powers that be prefer human censors. It is all about the who-whom question. Folks with the wrong politics are censored heavily while folks with the “right” (left) politics and opinions are allowed to talk about the same topics as much as they want. An A.I. would start censoring people on the left and that isn’t what they want.
For example, Hillary’s book said something like, ‘tragically there are some people who actually believe that whites and blacks have different IQs.’ An AI would censor her book because black, white and IQ appear together. TPTB wouldn’t want that.
But human censorship is bad enough.
Dan
September 21, 2017 at 11:36 AM
> tragically there are some people who actually believe that whites and blacks have different IQs
As Steve Sailer points out the last 72 finalists in the Olympic 100 meter dash have been of West African descent.
ScarletNumber
September 21, 2017 at 1:25 PM
Today we live in a bubble of free speech, but it’s a temporary bubble.
We live in a bubble of mostly garbage, lies, obscenities and general, low class stupidity.
I’m all for censorship. Let’s stop pretending the average moron out there shouldn’t be censored.
It’s just a question of who is doing it.
I try to censor all the time. I don’t watch TV, movies or listen to current music.
I try to be selective and censor as much as possible. But of course I can’t censor in some cases until it’s too late and then I am confronted with garage from dopes who can type or use a camera.
Most media content is of no value or significance to intelligent people.
Rifleman
September 21, 2017 at 2:42 PM
“We live in a bubble of mostly garbage, lies, obscenities and general, low class stupidity”
ProleMerica, ProleMerica,
oh garbage, lies
and travesty,
with morons from
sea to shining sea!
JS
September 21, 2017 at 8:13 PM
“I try to censor all the time. I don’t watch TV, movies or listen to current music.
I try to be selective and censor as much as possible.”
Same here, it’s 99% brain pollution. But you mean that we are discriminate in the media we consume. Censorship means silencing someone, and I’ve never had that power: the garbage we avoid is readily available to people with lower, less discriminating standards.
hard9bf
September 22, 2017 at 3:58 PM
As a software developer, I’m afraid you’re wrong; current AI is very, very good at telling what you’re talking about. That’s more or less what Google search does—or how Spotify can find music you like even better than you yourself can, how your email service can filter out all the spam, phishing and viruses, etc, etc.
snorlaxwp
September 21, 2017 at 4:05 PM
I don’t see how Spotify finding music you like is relevant here. It is about classification of music genres and maybe looking at what other people who liked the same artists chose to listen to. It is not like understanding speech. Even google doesn’t really have to understand speech in the context of conveying an idea, they need to only understand it in the context of giving an instruction, which is easier than understanding meaning.
Hashed
September 22, 2017 at 6:02 PM
One blog suggested using “google” in place of the N-word. Good idea.
sestamibi
September 21, 2017 at 9:52 PM
you mean (((google)))
Hashed
September 23, 2017 at 5:08 PM
I think it’s difficult to predict how things will turn out. For example, AI will make it extremely inexpensive to construct and operate your own version of Facebook, Youtube, Gmail, etc. So there will be a lot of competitors who will be able to say “We don’t censor your e-mail.”
Already, even without AI, Duckduckgo is a very good alternative to Google.
fortaleza84
September 21, 2017 at 11:53 AM
When does the google algorithm become public property?
It probably wouldn’t be that hard to nationalize it in the name of the public good if things ever really came down to it. If the search algorithm were nationalized, that’d probably be the end of google’s market cap.
You could force google to take a bailout (there is precedent for coercing private institutions to accept government funds) thereby making it subject to more government regulation.
Right now it’s more important to win hearts and minds so such drastic action would need to be avoided. If the left ever made any headway to truly censoring the right though, the gloves would have to come off.
Anybody can set up a mail server these days, so censoring email wouldn’t be a big deal.
Paul Ryan's Sickly Old Lap Dog
September 21, 2017 at 3:22 PM
“Anybody can set up a mail server these days, so censoring email wouldn’t be a big deal.”
Email without spam protection is pretty worthless.
The vast majority use Gmail or Microsoft free email services for personal use.
If you set up your own email server, but can’t send emails to anyone on Gmail or Microsoft because their AI has determined that your server is “racist,” then what good is it?
Lion of the Blogosphere
September 21, 2017 at 3:41 PM
The google pagerank algorithm is worthless. No one uses hypertext links anymore, except on right wing blogs. And not this one.
Helmut
September 23, 2017 at 11:07 AM
Facebook, YouTube, and Google are monopolies because of network effects and economies of scale. AI doesn’t change that. I use DuckDuckGo but Google is better for long tail searches.
JW Bell
September 21, 2017 at 3:39 PM
Google has BILLIONS of dollars to spend on improving its algorithnms, no startup can compete against that. Even Microsoft, another very wealthy company, can’t compete against Google and they’ve been trying for a long time.
Lion of the Blogosphere
September 21, 2017 at 3:48 PM
Microsoft can’t compete because it doesn’t compete and doesn’t know how to compete. Microsoft is very good at milking monopolistic products for all they are worth, but as for developing anything groundbreaking, or going toe to toe against an established player, forget about it. Microsoft’s long term strategy has been to buy it’s competitors. Can’t do that forever.
Microsoft should spin off Bing with a bundle of cash. First order of business for Bing should then be to charge nothing for search based advertising space for, say, 6 months, and gradually introduce fees, always keeping the fees significantly below Google’s, and always researching and developing ways to bring good bang for the buck to the customer (not the user). The efficacy of the money spent on Google advertising is being called into question. There is a story about Restoration Hardware and their discovery that they weren’t getting any value of the millions spent on Google. Other’s are probably reaching the same conclusion.
Bing still might fail but at least they would have put up a good fight. Presently, Microsoft does nothing with it and if it doesn’t do anything with it, Bing will eventually wither away as so many other Microsoft products have.
Daniel
September 21, 2017 at 5:30 PM
That reminds me of a couple of articles I ran across last week that I think tie in with each other as well as your post. Not that I’m criticizing AI. I for one welcome our new robot overlords.
https://www.theguardian.com/technology/2017/sep/12/artificial-intelligence-face-recognition-michal-kosinski
http://www.latimes.com/books/la-ca-jc-world-without-mind-20170912-story.html
destructure
September 21, 2017 at 12:30 PM
Here’s another on the same theme from one of the pioneers of AI.
https://www.axios.com/artificial-intelligence-pioneer-calls-for-the-breakup-of-big-tech-2487483705.html
destructure
September 21, 2017 at 1:39 PM
They’re going to shut down AI for being racist. Pattern recognition is racism.
everybodyhatesscott
September 21, 2017 at 1:27 PM
This is an ongoing cat and mouse game Lion, the moment ISPs deploy censorship technologies, users collectively adapt by deploying better technologies to become covert. Censorship by the ruling classes is older than the hills, as is the human ingenuity to beat it.
Roli
September 21, 2017 at 2:33 PM
Sherman anti-trust act.
Curle
September 21, 2017 at 2:44 PM
AI is still just a glorified routine machine that can’t think outside its programming.
That’s why automated customer service programs end up with the customer yelling “associate” into the phone.
The Undiscovered Jew
September 21, 2017 at 3:10 PM
I wonder why guys like Musk get all worked up over AI, but not ART.
gothamette
September 22, 2017 at 12:22 PM
The infiltration of Christian lunatics into the Air Force has gone so deep that even Latinos are getting involved: https://www.dailykos.com/stories/2017/9/21/1700384/-Air-Force-Standing-Behind-Chaplain-Who-Said-Obeying-the-Constitution-Serves-Satan
Bannon was right: we can’t destroy the stupid Christians, we need their help. Instead we should just exterminate the TruCons.
Otis the Sweaty
September 21, 2017 at 3:18 PM
The chaplain is right. There is always an Official State Religion. America’s should be Christianity.
Andrew E.
September 21, 2017 at 3:49 PM
I very carefully read the original article by Hernandez. He did not violate anyone’s religious rights or encourage others to do so. He simply told Christians not to violate their own religious beliefs. The lefties are just angry because they want to impose their leftist ideology and he’s pushing back.
destructure
September 21, 2017 at 5:54 PM
OT: Beta rage?
http://www.masslive.com/news/index.ssf/2017/09/orion_krauses_father_says_fami.html
Ripple Earthdevil
September 21, 2017 at 7:07 PM
What kind of jerk names his son Orion?
Richard
September 21, 2017 at 11:25 PM
All they need to do is design an algorithm to detect an ban any mention of the word “IQ”.
DataExplorer
September 21, 2017 at 8:44 PM
NeoGaf worried about the healthcare bill and Trump’s “rising” popularity.
HC isn’t gonna pass but the excuse the Gaffers are making for Trump’s apparent upsurge in popularity is that it is just some democrats who are approving of Trump because of DACA and that Trump’s numbers aren’t improving with indies. Sure, whatever.
According to PPP, Hillary beats Trump by 4 in a hypothetical election held today. They had Hillary beating Trump by 4 in the general, although with fewer undecideds then their current poll has.
My guess? If the election was redone today, we would get a tied popular vote.
Otis the Sweaty
September 21, 2017 at 9:08 PM
My guess? If the election was redone today, we would get a tied popular vote.
And in the electoral college, what?
Nobody cares if Trump loses New York and California by EVEN MORE.
This hurricane is going to push even more Puerto Ricans into the US, maybe into Florida and PA. Could effect the vote in 2020.
Rifleman
September 22, 2017 at 4:18 PM
Trump would win the electoral college by even more. Would add NH and MN.
The PR’s will hopefully mostly go to NY, but yeah if a lot of them go to PA or FL then we could have a problem. The good news is that after power is restored, we will probably only see a net migration of 250000 from the Island. Split 3 ways between PA, NY and FL that comes out to 80k per state, which translates into about a net 20k D votes each in those 3 states. Too small to make a difference.
Otis the Sweaty
September 22, 2017 at 5:05 PM
Now NeoGaf is saying that the have an opportunity for a Senate pickup in AL if Moore wins the primary. These people never learn.
Otis the Sweaty
September 21, 2017 at 9:19 PM
You should go on there and build up their hopes about a Moore defeat.
And then watch as they are crushed again.
Rifleman
September 22, 2017 at 4:19 PM
I can’t go on there. You need a paid email address to even apply for membership. Can’t use gmail or whatever. On top of that, the NeoGaf boss “evilore” is so paranoid about people trolling him that it takes upwards of a year to get an account approved.
They don’t need my help to get their hopes up for a Jones win, they are doing that all by themselves. DKos too.
Otis the Sweaty
September 22, 2017 at 4:55 PM
Hey Lion, how about a post on Bitcoin? The last you wrote was a comment in 2014
“Once a major government officially classifies it a Ponzi scheme and makes it illegal to sell it, that will be the end of bitcoin.”
Has your opinion changed?
Abe
September 21, 2017 at 11:54 PM
Looks like I was wrong. Although there continue to be news stories of people having their bitcoins savings stolen by hackers.
Lion of the Blogosphere
September 22, 2017 at 7:16 AM
Bitcoin makes for fun speculation and if you buy and sell at the right prices you can make a lot of money (in dollars of course). But it’s just gambling. Remember, Bitcoin is supposed to facilitate anonymous peer-to-peer payments. Why would anyone use Bitcoin to pay for something (something mundane, most people aren’t arch criminals) when the value could go from $3,000 to $4,000 in a week? You’d just hold (or hodl) onto it for the chance at capital appreciation. No, it will be a fun rollercoaster ride while the $IMF system still exists but it will never become a currency or a money or a store of value.
And to the extent that the public really wants a distributed ledger digital cryptocurrency for the convenience and the anonymity that they would provide (assuming the technology can be worked out) then the central banks will just adopt the tech and offer their own cryptos as part of their base money supply as an alternative to cash money at the ATM. See here for instance:
https://www.bis.org/publ/qtrpdf/r_qt1709f.htm
Andrew E.
September 22, 2017 at 9:22 AM
An examination of how Trump’s inevitable firing of Mueller will play out: http://www.latimes.com/opinion/op-ed/la-oe-litman-mueller-firing-20170921-story.html
Otis the Sweaty
September 22, 2017 at 12:31 AM
OT — The comments in one of Sailer’s recent posts are right up your alley: http://www.unz.com/isteve/silicon-valleys-most-boring-adventuress-is-back/
ipc
September 22, 2017 at 2:15 AM
The N-word is uttered by blacks in 99.9% of cases where it is heard in America today.
Social media will not be able to deprive blacks of their favorite term.
Camlost
September 22, 2017 at 2:28 AM
Blacks say “nigga” and not the bad n-word.
Lion of the Blogosphere
September 22, 2017 at 7:16 AM
I know, but social media would have to allow one to stay and censor the other. That would be hard to defend.
Camlost
September 22, 2017 at 7:50 AM
That’s the excuse given by blacks and SJW’s, but then according to those same people, non-blacks aren’t allowed to say “nigga” either. Here’s a tweet my left-wing cousin’s Twitter feed led me to expressing that sentiment:
Hermes
September 22, 2017 at 12:12 PM
Comment from Dkos about Congressional Dem’s resistance to Obamacare repeal: “Where are the Dems? Why aren’t they out there screaming and shouting the injustice of this — no they are more concerned about Dreamers. Frankly I care more about healthcare for all than illegals who feel they are entitled to stay here.
Dems take care of your supporters/voters first -we are your citizens and we vote!”
Comment has 2 upvotes.
Otis the Sweaty
September 22, 2017 at 3:02 AM
Washington Post is so happy to cite Kim insult , they have it on front page since 24 hours. WP really hates Trump even more tha NYT . I don’t know why ….
Bruno
September 22, 2017 at 6:21 AM
AI is just pattern recognition and the most ambitious AI projects that have been unleashed in to the wild were all shut down because AI is so virulently racist. The only way to stop AI from being racist is to manually brick it in regards to a whole host of topics which produces a less racist AI, but also a stupider one that will be out thought by China’s non-bricked AI machines.
PerezHBD
September 22, 2017 at 7:50 AM
O/t — but related to Trump’s aides caught in the Russia scandal:
http://www.msn.com/en-us/news/politics/for-trump-aides-caught-in-russia-probe-legal-bills-and-paranoia/ar-AAsfq6R?li=BBmkt5R&ocid=ientp
Here’s a pic of the guido, Mike Caputo, a Trump aide purportedly at his home, preparing his testimony, while wearing his pj-T-shirt that reads “Dead”.
What an effing moron!
JS
September 23, 2017 at 1:18 PM
There is a reason “it” is called “artificial.”
thordaddy
September 23, 2017 at 4:40 PM