View profile

Fresh battles over borderline content

October 15 · Issue #400 · View online
The Interface
Our first-ever live event takes place in San Francisco on Tuesday, October 22nd: I’ll be talking with the brilliant disinformation researcher Renee DiResta. Grab a ticket and come say hi!
Over the past week or so, it has felt like the entire public conversation around tech has turned a now familiar question: what should stay up, and what should come down?
First, there’s the ongoing debate about whether Facebook should permit politicians to lie in their ads. (A tiny bit more on that below.)
Second, a pro-Trump conference screened a video of footage from the 2014 film Kingsman: The Secret Service edited to depict “the president shooting, stabbing, and attacking the news media and his opponents.” YouTube restricted the video’s audience but did not take it down; Twitter briefly suspended a meme maker associated with the group who made the video, over a copyright violation.
Third, Twitter sought to clarify its never-enforced rules around what to do if a “world leader” ever incites violence against a private individual, promotes terrorism, or violates a handful of other big rules. In some cases it will remove the tweet; in others it will put it behind a warning. That’s all well and good, but policy is what you enforce — and until Twitter takes an action on one of these tweets, it’s really just talk.
Fourth, Googlers are mad that the company removed The Revolution of Our Times, a game that “allows players to participate in virtual recreations of the pro-democracy demonstrations” in Hong Kong. Google says it’s part of a “long-standing policy” against creating games that take advantage of “sensitive events,” but some employees smell a rat. 
Fifth, TikTok — owned by the Chinese company ByteDance — announced it had hired and outside law firm and formed a committee to refine its content moderation guidelines and disclose them more fully.
Finally … oh god, how did Airbnb get mixed up in all this? Here’s Tonya Riley:
Fair-housing advocates are accusing Airbnb and other short-term rental platforms of abusing the legal provision that grants tech companies broad immunity for content people post on its platforms. They want Congress to ensure online rental services cannot ignore — and profit off— listings that violate state and local housing laws. 
Most of the debate over Congress reviewing of Section 230 of the Communications Decency Act has focused on social media companies: Republicans argue they’re using the law to unfairly silence conservative voices and Democrats say the act should be changed to ensure social media giants take greater responsibility for disinformation and other problematic content.
But this latest attack, coming just days before a House Energy and Commerce joint subcommittee hearing to examine the decades-old law, shows that changes could have consequences for other tech platforms such as online marketplaces. 
OK, so that one is kind of an outlier … or at least, another conversation.
As for the rest, it strikes me that they all revolve around what platforms sometimes refer to, euphemistically, as “borderline content.” The basic idea is that the closer a piece of content comes to violating a platform’s rules, the more that people tend to engage with it. In a world where many people want more attention than they’re getting, creating borderline content is one of the few proven recipes for success.
But because it may violate a platform’s rules, it requires a lot of agonizing discussions among members of the policy team and other executives. And because borderline content tends to touch on sensitive or controversial subjects, the backdrop of those discussions is often a public-relations crisis.
What to do about it? Platforms have recently said that they hope to make borderline content less prominent in their feeds. Facebook said it would de-emphasize sensationalist stuff last November; YouTube followed suit in June. Twitter will probably have something to say about all this in 2022.
But even if that content is less prominent to most users, it will still be hosted on company servers — and will require intervention from moderators. I expect that to be doubly true next year in the United States, when we have a presidential election to look forward to.
In short, the swarm of headlines about content moderation over the past week should not be mistaken as a coincidence. What stays up — and what comes down — has never been a more salient question in people’s minds. And absent any meaningful regulation, expect the tech platforms to keep fumbling their way forward, trying to appease as many users as possible.

Thanks to everyone who wrote in with thoughts about Facebook, political ads, and lies. I’ve already said a lot on this subject, but I wanted to highlight two popular ideas in your replies. One is that lying in political ads might not feel so noxious if politicians had to lie to Facebook’s entire user base, rather than micro-targeted segments — where their lies might go unnoticed. One of you wrote:
Could there be a partial solution for disinformation that won’t abridge our free speech rights by banning micro targeting? What if instead of trying to constrain what people say in ads we force them to say it to everyone? Ban micro targeting for all political ads. Make a law that all political ads must be broadcast to all interests in the region affected. So national election ads must broadcast nationally regardless of party affiliation or interest. State election ads must broadcast state wide etc. 
More radically what if internet ads had to be paired with message equivalent TV, cable, and newspaper ads? I bet that would reduce the effectiveness of disinformation campaigns because it would make it harder (but not impossible) to radicalize the target audience without inflaming the opposition. 
Another popular idea, and one that Facebook has considered in the past: just ban political ads altogether. This feels overly defeatist to me — and could disadvantage good candidates who don’t yet have national name recognition. But given the public-relations beatings that I expect Facebook to take over its new policy in the coming year, I wouldn’t be surprised if the company eventually did ban political ads.
I also liked this point, made in an op-ed from Ellen Goodman and Karen Kornbluh in the Guardian, about the idea that not fact-checking political ads will make the company “neutral” politically. (Emphasis mine.)
Facebook seems to concede that it — like broadcasters — exercises gatekeeping control over attention, advertising dollars, and political debate, and therefore has a fiduciary responsibility of some kind. But the platform wants to cherrypick only the permissive aspects of regulation: don’t moderate for disinformation. What Facebook fails to acknowledge is that it isn’t neutral. It is favoring candidates who smear their opponents and amplify baseless conspiracies. It’s not just that the platform takes these ads; its algorithmic design juices their circulation by advantaging the incendiary over the informative to increase engagement.
Elsewhere: BuzzFeed looks at the political ads Facebook does ban, such as showing a picture of a button that you can’t actually click.
The Ratio
Today in news that could affect public perception of the big tech platforms.
🔼 Trending up: Here’s some cautious optimism for TikTok’s creation of a new committee to help shape its content moderation policies, with help from the corporate law firm K&L Gates. The company has come under fire recently for censoring content related to the Hong Kong protests while allowing the #trump2020 hashtag to thrive.
🔽 Trending down: Thanks to its decision to ban a game that lets you recreate the Hong Kong protests, Google has to set its “days without a major employee protest” counter back to zero.
The 19 Democrats running for president have poured nearly $32 million into Facebook ads this year. And we can now see which voters they’re targeting — and how — thanks to Facebook’s public ad library. Here’s Shane Goldmacher and Quoctrung Bui at The New York Times:
When it comes to former Vice President Joseph R. Biden Jr.’s Facebook ads, the odds are overwhelming that people seeing them were born before 1975. For Senator Bernie Sanders of Vermont, the opposite is true: People born in 1975 or after are more than twice as likely to see his Facebook ads than those born earlier. A gender split is clear, too: About half of Mr. Sanders’s audience are men, while about two-thirds of Mr. Biden’s are women.
Senator Elizabeth Warren of Massachusetts, more or less, splits the difference between them. She has targeted a Facebook audience that more closely mirrors the projected universe of likely Democratic primary voters in 2020: more women than men (though not by as much as Mr. Biden), leaning older more than younger (but again, not by as much as Mr. Biden).
She might be spending big on Facebook ads. But Elizabeth Warren says she’s no longer accepting large donations from Facebook and Google, and wants other candidates to do the same. The new policy is part of Warren’s plan to get big money out of politics — something she is expected to hit on during tonight’s debate. (Makena Kelly / The Verge)
Algorithms are transforming how government aid is distributed in countries around the world. Now, instead of case workers fighting for people’s food and housing subsidies, data plays a key role in deciding who gets what. The results are generally worse for people with less money. (Ed Pilkington / The Guardian)
Thanks to a single liked tweet, Tencent apparently told ESPN that it will no longer broadcast Adrian Wojnarowski’s Chinese basketball show, Woj in the House. Wojnarowski made the fateful mistake of liking Daryl Morey’s now infamous pro-Hong Kong tweet, which China did not appreciate. (Laura Wagner / Deadspin)
The architect-turned-whistleblower of the Cambridge Analytica data privacy scandal, Christopher Wylie, has a new book out. It’s part memoir, part exposé, and tells the story of how an Obama staffer got swept up in a dark web of data manipulation in support of candidate Trump. (Steven Levy / Wired)
Amazon sellers are using Facebook chatbots to strong-arm people into leaving good reviews. Here’s how it works: sellers post Facebook ads touting “free” products. When someone clicks on the ad, they’re told to message the company. When they do, a bot answers telling them they’ll get a full refund after buying the product, but only if they leave a 5-star review. Here’s more on the scheme from Nicole Nguyen at BuzzFeed:
Some Amazon sellers use a chatbot to automate the entire interaction. “Hello! Thanks for your interest!” responds one Facebook page, named VPOW GE. “Do you wanna test our FREE portable breathalyzer? It’s [sic] real cost is USD 33.99, but now it is free for testing in limited time! (If yes, simply click the “yes” button under the picture to proceed.)”
Each of the page’s questions were followed by clickable “yes” and “no” buttons. Do you have Amazon and PayPal accounts? Yes. Can you leave a review one week after receiving the package? Yes. Confirm that your first review was published before June 1, 2019. Yes, confirmed.
The page then offers a detailed set of instructions. Search for a particular keyword, look for the brand name WEIO, add the product to your Amazon Wish list, then order the product. Make sure you don’t use a gift card. Half of the refund will be issued within 24 hours of receipt. The other half will arrive after you send a screenshot of your review. But be sure to wait ONE WEEK before reviewing.
TikTok recently moved to a new office in Mountain View that previously housed Facebook-owned WhatsApp. Since 2018, the company has poached more than two dozen Facebook employees, by paying them up to 20 percent more. Now that they’re in Facebook’s backyard, that number could grow larger. (Salvador Rodriguez / CNBC)
Meanwhile, Facebook is expanding in Seattle. The Pacific Northwest office now employs more than 5,000 people, making it the second-largest Facebook workspace outside of the Menlo Park headquarters. (Monica Nickelsburg / GeekWire)
Reviews of Facebook’s new Portal are out. Our colleague Dan Seifert at The Verge say the device offers “better than video calls on the smart displays from Amazon or Google.” Bloomberg called the device a “stellar video chat camera, weighed down by privacy concerns and better competition.” This is all basically what critics said last year.
Kids are outsmarting Apple’s parental controls meant to limit their screen time by exploiting software bugs and hacking into their parent’s accounts. The feature, introduced last year in response to Tristan Harris’ Time Well Spent movement, was supposed to help counteract tech addiction. But Apple has been slow to close the most glaring loopholes. (Reed Albergotti / The Washington Post)
Twitter’s head of product, Kayvon Beykpour, came on The Vergecast to discuss upcoming features, how the company is handling verification, and whether or not they’ll ever roll out an auto-delete feature for tweets. Casey co-hosted the interview alongside our boss, Nilay Patel — I hope you’ll check it out. (Andrew Marino / The Verge)
A website called YouTube Decade lets you see the videos that have gotten the most views since the day they were uploaded 10 years ago. If you’re feeling nostalgic for slightly simpler internet-times, this could do the trick. (Jay Peters / The Verge)
Recode launched a new podcast called Reset about how tech is changing our lives. It’s hosted by our friend Arielle Duhaime-Ross, who was formerly the first climate change correspondent in American nightly TV news. Reset comes out thrice weekly; the debut episode is about bio-hacking. (Liz Nelson / Recode)
And finally ...
Just as never-ending social feeds sent some of us running back to email, so too are never-ending direct messages sending people back to text messages. The latest example, via Taylor Lorenz: celebrities encouraging you to text “them” directly. For the most part, this seems to just sign folks up for a promotional SMS mailing list.
On the other hand, you now may or may not have access to the artist formerly known as Puff Daddy:
“I’m sitting here and I’m thinking, I’m about to go into this next era of my life and I’m going to be doing a lot of positive things, a lot of disruptive things, a lot of things I really don’t want everybody, like everybody to know about,” Diddy said in an IGTV video posted on Oct. 2. “On the ’Gram, everybody knows about everything. I want a deeper connection with my fans.”
His number is 917-746-1444. Let us know if you get through!
Talk to us
Send us tips, comments, questions, and problematic content: and
Did you enjoy this issue?
In order to unsubscribe, click here.
If you were forwarded this newsletter and you like it, you can subscribe here.
Powered by Revue