AI not working as intended

ft_19.07.11_socialmediacompanies_younger-republican-men-less-likely-see-need-remove-offensive-content_2
source

As we have previously covered, there has been a push in recent times for social media companies to censor posts containing offensive remarks, Fake News, Russian bots, or slowed down videos of Nancy Pelosi. Social media companies have been dumb enough to along with this demand (mostly because the people who run them are also the kinds of people who think social media companies should censor more,) but as a practical matter, they are of course incapable of reading and evaluating every post that comes their way.

So these companies need algorithms to trawl their millions of posts for offensive content.

Unfortunately (for someone) any AI trained on things that people actually find offensive will censor things the Woke don’t want censored:

On LGBTQ pride month, we share some of the main findings of our research looking at the impacts of Artificial Intelligence on LGBTQ speech. Our goal is to shed some light on the gaps and biases that may be present in AI technologies that are currently being developed to moderate content on internet platforms and demonstrate how they might have significant implications for LGBTQ rights.

The results of this study are unintentionally hilarious:

By training their algorithm to learn what pieces of content are more likely to be considered as “toxic”, Perspective may be a useful tool to make automated decisions about what should stay and what should be taken down from the internet platforms. …

We used Perspective’s API to measure the perceived levels of toxicity of prominent drag queens in the United States and compared them with the perceived levels of toxicity of other prominent Twitter users in the US, especially far-right figures. …

After getting access to Twitter’s API, we collected tweets of all former participants of RuPaul’s Drag Race (seasons 1 to 10) who have verified accounts on Twitter and who post in English, amounting to 80 drag queen Twitter profiles.

We used Perspective’s production version 6 dealing with “toxicity”. We only used content posted in English, so tweets in other languages were excluded. We also collected tweets of prominent non-LGBTQ people (Michelle Obama, Donald Trump, David DukeRichard SpencerStefan Molyneux and Faith Goldy). These Twitter accounts were chosen as control examples for less controversial or “healthy” speech (Michele Obama) and for extremely controversial or “very toxic” speech (Donald Trump, David DukeRichard SpencerStefan Molyneux and Faith Goldy). In total, we collected 116,988 tweets and analysed 114,204 (after exclusions).

grafico_com_drags
“drag queens are in blue; white supremacists in red; Michelle Obama in green and Donald Trump, in orange.”

It turns out that drag queens are really rude and Richard Spencer is pretty polite. Even Donald Trump, who is kind of rude sometimes, is more polite than the majority of drag queens.

Of course, anyone who has actually read their tweets would have known this already. As the study notes, drag queens are likely to pepper their tweets with phrases like “I love you bitch” and “I’m a faggot,” while white supremecists are likely to say things like “Italy has a lovely history and culture and we should preserve it.” Drag queens also know their speech is protected (back in the early days of systems like AOL Online, you weren’t allowed to use words like “bitch” and “fuck,” but today they are allowed,) while WNs know that they are being watched and that if they step out of line, there’s a good chance their accounts will be deleted.

Any algorithm trained on actual rudeness as perceived by normal humans will of course tag drag queens’ typical tweets as offensive; it will take intervention by actual humans to train the algorithms to pick up on the words creators actually want to censor, like “Donald Trump” or “Chinese detention camps”.

 Speaking of which:

Van de Weghe has continued to study Chinese AI—how it tracks people with ever-improving facial recognition software. He describes the new “social credit” programs that use AI to combine data from numerous sources, assign scores to people’s behavior and allocate privileges accordingly. In 2013, when Liu Hu, a Chinese journalist, exposed a government official’s corruption, he lost his social credit and could no longer buy plane tickets or property, take out loans, or travel on certain train lines. …

Jennifer Pan, an assistant professor of communication, explains why Chinese citizens accept social credit programs. “People think others spit in the street or don’t take care of shared, public facilities. They imagine that social credit could lead to a better, more modern China. This is an appealing idea. Political dissent is already so highly suppressed and marginalized that the addition of AI is unlikely to have anything more than an incremental effect.”

The result for journalists is that actual prisons (where many are currently held) are replaced by virtual prisons—less visible and therefore more difficult to report on. In the face of this, Van de Weghe says, many journalists he knows have quit or self-censored. And while reporters outside China can critique the general practice of censorship, thousands of individual cases go unnoticed. Government computers scan the internet for all types of dissidence, from unauthorized journalism to pro-democracy writing to photos of Winnie-the-Pooh posted by citizens to critique President Xi Jinping, who is thought to bear a resemblance. AI news anchors—simulations that resemble humans on-screen—deliver news 24/7. The government calls this media control “harmonization.” The Communist Party’s goal for sustaining its rule, according to Pan, “is to indoctrinate people to agree. Authoritarian regimes don’t want fear.”

There is a lot of amazing technological progress coming out of China these days, for good or bad:

If you think mass AI-censorship and surveillance sounds scary in China but totally good and awesome in the US, you haven’t thought this through.

Algorithmic Optimization pt 2

 

There is nothing exceptional about the slowed-down Nancy Pelosi video, and nothing terribly exceptional in reporters saying uninformed things about subjects they aren’t well versed in.

The significance lies far more behind the scenes. From Marketwatch: Facebook Decided to Rethink Policies on Doctored Media two days Before Pelosi Video.

Wow, that is awfully coincidental that Facebook just happened to be thinking about changing these policies anyway right before a doctored video just happened to make it onto the news, prompting millions of people to pressure Facebook into doing exactly what Facebook already wanted to do.

Don’t be fooled: this isn’t spontaneous. Oh, sure, many of the people at the low end, like reporters, are just doing their job of reading the news they have been given into the camera, but there is plenty of active coordination going on behind the scenes by organizations like Facebook and the Democratic Party.

The Democrats realized sometime around 2016 that they have a meme problem. People on the internet thought Trump was funny and Democrats were boring sticks in the mud. People on the internet made videos about Hillary Clinton’s health, the European migration crisis, and other subjects the Dems didn’t approve of.

They don’t want this happening again.

So they are laying the groundwork now to re-write the policies and algorithms to strategically remove problematic conservative voices from the fray. Alex Jones has already been kicked off Youtube, Facebook, PayPal, etc. FB has taken a particularly hard line, threatening not just to delete Jones’s videos, but any account that posts them (excepting those that post them in order to criticize them).

Even Visa and Mastercard are getting in on the act, cutting off banking services to organizations whose political views they don’t like.

The ostensible reason for Alex Jones’s deplatforming is his supposed spread of conspiracy theories post-Sandy Hook (I say “supposedly” because I have not seen the clips in question,) but it is obvious that 1. these concerns surfaced years after Sandy Hook and 2. no one has deplatformed media outlets that pushed the “Iraq has WMDs” conspiracy theory that cost the US trillions of dollars and lead to the deaths of thousands (millions?) of people.

This has all been accompanied by a basic shift in how media platforms and infrastructure are viewed.

The traditional conception is that these are platforms, not publishers, and thus they merely provide something akin to infrastructure without much say over how you, the user, put it to use. For example, the electric company provides electricity to anyone who pays for it, and even if you use your electricity to warm the cages of your illegally gotten, exotic, endangered reptile collection, the electric company will generally keep providing you with electricity. The electric company does not have to approve of what you do with the electricity you buy, and if you break the law with their electricity, they see it as the state’s job to stop you.

A publisher and a platform, like Facebook, traditionally enjoyed different legal rights and safeguards. A publisher checks and decides to publish every single item they put out, and so is held to be responsible for anything they print. A platform merely provides a space where other people can publish their own works, without supervision. Platforms do not check posts before they go up, (as a practical matter, they can’t,) and thus are generally only held legally responsible for taking down material on their site if someone has notified them that it is in violation of some law.

EG, suppose someone posts something really illegal, child porn, on Facebook. If Facebook is a “publisher,” it is now publishing child porn and is in big legal trouble. But since Facebook is just a platform, it deletes the videos and is legally in the clear. (The poster may still go to prison, of course.)

The conceptual shift in recent years has been to portray platforms as “allowing” people to come in and use their platforms, and then ask why they are allowing such shitty people to use their platform. No one asks why the electric company allows you to use their electricity to raise your army of bio mutant squids, but they do ask why Facebook allows right-wingers to be on the platform at all.

This is treating platforms like publishers, and they are absolutely jumping into it with both feet.

Let’s skip forward a bit in the video to the lady in white to see this in action:

It’s been viewed millions of times on the internet, but it’s not real… This is really scary, and not going away, and I’m fearful this is going to be all over the 2020 election.

You know, that’s how I felt when libs kept bringing up Harry Potter in the context of the last election, but for some reason taking a children’s fantasy story about wizards is acceptable in political discourse but slowed-down videos aren’t.

And who is responsible for monitoring this stuff, taking it down? Facebook, Youtube took it down, but after how long?

Other commentator… At Facebook it’s still up because Facebook allows you to do a mock video…

The correct answer is that no one is responsible for monitoring all of Facebook and Youtube’s content, because that’s impossible to do and because Facebook isn’t your mommy. If you want Facebook to be your mom and monitor everything you consume, just stop talking and leave the adults alone.

CNN asks:

So Monika, in the wake of the 2016 election obviously Facebook has repeatedly told Congress and the American people that yo’re serious about fighting disinformation, fake news, and yet this doctored video which I think even your own fact checkers acknowledge is doctored of speaker Pelosi remains on your platform. Why?

Like the previous guy already said, because it’s not against Facebook’s TOS. Of course Anderson Cooper already knows this. He doesn’t need to get an actual Facebook representative on his show to find out that “funny reaction videos are allowed on Facebook.” And if Facebook were serious about maintaining its neutrality as a platform, not a publisher, it would not have bothered to send anyone to CNN–it would have just left matters at a blanket statement that the video does not violate the TOS.

The Facebook Lady (Monika,) then explains how Facebook uses its algorithms to demote and demonitize content the “experts” claim is false. They’re proud of this and want you to know about it.

So misinformation that doesn’t promote violence, but misinformation that portrays the third most powerful politician in the country as a drunk or as somehow impaired, that’s fine? 

Oh no, quick, someone save the third most powerful person in the country from people saying mean things about her on the internet! We can’t have those disgusting peasants being rude to their betters!

Anderson Cooper is infuriatingly moronic; does not “logically understand” why Facebook leaves up videos that don’t violate the TOS but suggests that Facebook should “get out of the news business” if it can’t do it well.

Facebook isn’t in the “news business” you moron, because Facebook is a platform, not a publisher. You’re in the news business, so you really ought to know the difference.

If you don’ know the difference between Facebook and a news organization, maybe you shouldn’t be in the news business.

That said, of course Anderson Cooper actually understands how Facebook works. This whole thing is a charade to give Facebook cover for changing its policies under the excuse of “there was public outrage, so we had to.” It’s an old scam.

So to summarize:

  1. The Dems want to change the algorithms to favor themselves, eg, Facebook decided to rethink policies on doctored media two days before Pelosi video, but don’t want to be so obvious about it the Republicans fight back
  2. Wait for a convenient excuse, like a slowed-down video, then go into overdrive to convince you that Democracy is Seriously Threatened by Slow Videos, eg, Doctored Videos show Facebook willing enables of the Russians; Doctored Pelosi video is leading tip in coming disinformation battle
  3. Convince Republican leadership to go along with it because, honestly, they’re morons: Congress investigating deepfakes after doctored Pelosi video, report says
  4. Deplatform their enemies
  5. Rinse and repeat: Vox Adpocalypse

 

One final note: even though I think there is coordinated activity at the top/behind the scenes at tech companies and the like, I don’t think the average talking head you see on TV is in on it. Conspiracies like that are too hard to pull off; rather, humans naturally cooperate and coordinate their behavior because they want to work together, signal high social status, keep their jobs, etc.

Cathedral Round-Up #15: Duke

Duke literally looks like a cathedral
Duke literally looks like a cathedral

For this month’s Cathedral Round-Up, I decided to look beyond the Ivies at Duke University, North Carolina. For my non-American readers unfamiliar with our less famous institutions, Wikipedia states:

Duke is the seventh-wealthiest private university in America with $11.4 billion in cash and investments in fiscal year 2014.[9]

Duke is consistently included among the best universities in the world by numerous university rankings,[10][11] and among the most innovative universities in the world.[12] According to a Forbes study, Duke is ranked 11th among universities that have produced billionaires.[13][14] In a New York Times corporate study, Duke’s graduates were shown to be among the most sought-after and valued in the world,[15] and Forbes magazine ranked Duke seventh in the world on its list of ‘power factories’ in 2012.[16]

Duke’s research expenditures in the 2014 fiscal year were $1.037 billion, the seventh largest in the nation.[17] In 2014, Thomson Reuters named 32 of Duke’s professors to its list of Highly Cited Researchers, making it fourth globally in terms of primary affiliations.[18] Duke also ranks fifth among national universities to have produced Rhodes, Marshall, Truman, Goldwater, and Udall Scholars.[19]Ten Nobel laureates and three Turing Award winners are affiliated with the university. Duke’s sports teams compete in the Atlantic Coast Conference and the basketball team is renowned for having won five NCAA Men’s Division I Basketball Championships, most recently in 2015.

It’s always good when your college is good at playing keep-away.

From Duke Magazine’s special all-language-articles edition:
The Power of Pronouns:

This past February I was invited to give a lecture at a Duke seminar called “LGBTQ Activism and History.”

[People declared their preference to be addressed with gender-neutral pronouns]

I’ll admit all this seems newish and complicated to me. It was only a few months ago that a neighbor’s teenage daughter explained to their parents and friends that they were “pan-sexual” (the sexual attraction to a person of any sex—male or female—or gender—masculine, feminine, or somewhere in between). Now the request was for us to use gender-neutral pronouns. We tried but flubbed it, until finally the thirteen-year-old exclaimed, “Please respect my pronouns” —and the light went on. What they were saying was, “Please respect who I am.”

Please stop having deep discussions with your neighbor’s barely pubescent teenage daughter about who she wants to bang. It makes you sound like a creep.

We don’t let 13 yr olds drive cars, work, live alone, vote, sign contracts, or have sex, because 13 yr olds are idiots who are still dependent on their parents to take care of them and keep them alive. Children should be treated with kindness and compassion, but we don’t respect their ideas on adult matters for the same reason we don’t let them live on their own.

The Places Words Go:

Intersectionality, a concept that started in academia and became popular among grassroots activists, recently has exploded in broader culture. … So when Hillary Clinton, on March 6, 2016, tweeted that: “We face a complex, intersectional set of challenges…” it signaled intersectionality’s full entrance into the mainstream.

Yet, what does it mean for this discourse, which originated in black feminist circles, to now enjoy popularity in a variety of contexts and uses? …

Clinton’s usage of intersectional language, while maybe well-intentioned, displays the slippage and de-radicalization that attends many popular uses. Intersectionality becomes a matter of drawing connections between multiple problems and multiple solutions. Losing sight of larger structural critiques of white supremacy, capitalism, and patriarchy, the problems become about discrimination and about a lack of opportunities or parity for various identities within our economic system. Instead of challenging neoliberal policies that prioritize privatization and investment, the market—by including everyone and improving the stakes of those already within it—becomes the foundation to break all barriers. …

Daniel José Camacho is a master’s of divinity student, pursuing ordination in the Presbyterian Church (U.S.A.).

In related news: Can an atheist lead a Protestant Church? A battle over Religion in Canada:

The Rev. Gretta Vosper is a dynamic, activist minister with a loyal following at her Protestant congregation in suburban Toronto. She is also an outspoken atheist.

“We don’t talk about God,” Vosper said in an interview, describing services at her West Hill United Church, adding that it’s time the church gave up on “the idolatry of a theistic god.” …
Ordained in 1993, the 58-year-old Vosper says she began questioning God’s existence 15 years ago and openly came out as an atheist in 2013.

Vosper said that the United Church has a tradition of “pushing the envelope” and pulling down barriers — in accepting the ordination of women, embracing the LGBT community and performing same-sex marriage. She said her views about religion have evolved. After initially rejecting the idea of a supernatural god and the idea of god as “the father,” she moved eventually to rejecting God completely. Instead, she preaches values, including justice, compassion and love. …

Like other mainstream denominations, the United Church of Canada, founded in 1925 as a merger of several denominations, has seen its numbers fall sharply in recent years. It reported having 436,292 members at the end of 2014, less than half the 1,063,951 it had at its peak in 1964.

Long story short, the church is struggling to remove her on the grounds that being an open atheist is kind of counter to the basic founding documents of the church:

“It’s tough on the United Church because we’ve created this mantra of inclusiveness and now it’s been tested. It goes against the grain to tell somebody that you have to leave.”

A Way to Protect All Ideas:

I attended two conferences, interviewed ten women, met another fifteen remarkable women, and produced twenty YouTube videos in eight short weeks just to answer one question: How do we address online hate speech while maintaining free speech?

Well, that certainly sounds like a randomized, large-N, unlikely to be biased sample of people.

I would like to think I began my research as an objective bystander. … As much as I hated the dangers women faced online, I also abhorred content-based censorship. I thought my desire to protect both women and speech online would ensure my objectivity.

Then the interviews began and my objectivity faltered. I listened as women told me how their ex-boyfriends non-consensually shared nude photos of them online; …

Conservatives have been telling women for years that it’s a bad idea to send naked pictures of themselves out into the world. Guess they were right.

I noticed a common theme to these stories: men using online hate and violence to silence women. I could barely fathom why hate speech intended to silence women was acceptable, but censorship of same hate speech is unacceptable. So I used my voice to speak against hate speech; I proudly declared myself a feminist in a YouTube video.

Unfortunately (and unsurprisingly), my declaration didn’t end the misogynistic speech. While the First Amendment guarantees protection from Congress silencing my feminist speech, there would be no guaranteed protection against a cyber-mob trying to silence me with rape and death threats. …

It’s not enough to protect freedom of speech from a government violation. We must also protect the freedom of speech of the disempowered from the empowered.

That’s not easy, but I realized that it is marginally easier when we speak together. That was the most rewarding part of my summer research: meeting all the women who, supported by their tight-knit community, courageously and collectively speak out against online hate.

Meanwhile in the real world:

Trump Supporter Beaten with a Crowbar:

Police in Northern New Jersey say a 62-year-old man was beaten with a crowbar outside a restaurant for wearing a T-shirt in support of Donald Trump’s presidential campaign. …

The victim was treated on the scene for injuries to his forearms, hands and thighs shortly after 6 p.m., the website reported.

“The motorist inquired why [the man] was wearing the shirt, directing profanities at him,” Bloomfield Police spokesman Ralph Marotti said, the New Jersey Star Ledger reported Tuesday. “The [victim] continued to walk away as [the] motorist followed him.”

Trump supporter sucker punched to the ground by Mexican at Trump Rally in San Jose:

I’ve yet to find any articles in Duke Magazine about respecting people’s right to walk in public without being viciously beaten for their political beliefs.