Daphne Keller Profile picture
Apr 1, 2021 27 tweets 7 min read Read on X
The meme that “platforms algorithmically amplify polarizing content because engagement drives ad revenue” has gotten seriously out of hand. Someone needed to burst that bubble. It’s such a bummer that the someone is Facebook VP Nick Clegg. 1/
nickclegg.medium.com/you-and-the-al…
People will give his points less credence because of who they’re coming from. (And posting them on Medium, as if they were independent musings and not a message crafted by Facebook’s Comms and Policy teams is an interesting branding gesture but isn’t going to fool anyone.) 2/
He’s saying so many things that are right, though. 3/
Not right as in “this means the ads + amplification meme has zero basis.” Right as in “this theory is only useful with major caveats and complications, so stop leaning on it like you’re a 17-year-old who just discovered Marx or Ayn Rand.” 4/
See, I’m almost as grumpy about this as Clegg is. But more at liberty to express myself. 5/
Did I mention I am also so, so close to finishing a big paper on this topic – focused on amplification and the First Amendment? I’ve thought about this a whole lot. 6/
To be clear, I think a lot of what FB says in the realm of platform regulation is cynical and self-interested. FB would thrive under regulations that pull up the ladder behind it, imposing burdens smaller competitors can’t shoulder. 7/
That's what we should see every time Clegg says this, which he says a lot 8/ Image
But two big things he says strike me as unavoidably true. First, a rational ad-revenue driven company will not want to ceaselessly amplify content that pushes our buttons in the short term but leaves us feeling icky in the long term. 9/ Image
The point about advertisers is real. Remember this: theguardian.com/technology/201… 10/
And this, which is not creepy at all 12/ wfanet.org/knowledge/item…
Platforms may want perfect the algorithm to amplify engaging content right up to the point before it triggers user and advertiser defection, of course. But that really complicates the ads+amplification meme. (Also it makes platforms a lot like regular media.) 13/
I do think Clegg gets carried away with this claim, and undermines the credibility of his valid points. 14/ Image
His second unambiguously true point is this. People respond to emotionally engaging content (and worse: bias-affirming, tribal-identity-reinforcing, and rage-inducing content) because we are hardwired to do that. It’s how our brains work. 15/ Image
Algorithms may detect from our behavior a revealed preference for sensational content, and give us more. But we also make that stuff go viral all by ourselves -- on platforms with no ads, and no algorithmic ranking. 16/ Image
A smaller point: I think Clegg (or his Comms people) miss a step here, in a way that will unnecessarily anger people from the news business in particular. Clickbait headlines evolved because people click on them, but also because they got algorithmic promotion juice. 17/ Image
It's kind of like platforms saying “we will reward this behavior until you adapt your business and get too good at it, and then we will punish it.” That’s the unavoidable cycle with spam or SEO. For news orgs, it has been awful. 18/
I’ll end on a note of agreement. Platform critics sometimes say, roughly, “if platforms can’t fix X, then they shouldn’t exist.” That’s fine as a moral stance. But the not-existing thing isn’t going to happen. So it’s a bad starting point for real-world analysis or advocacy. 19/ Image
OK, gotta get back to actually writing my amplification article. (It is SO CLOSE to being done.) If you’re interested in good work on this in the meantime I recommend @PJLeerssen papers.ssrn.com/sol3/papers.cf… and @ellanso et al ivir.nl/publicaties/do… 20/
Super sorry the link at the top of this is broken. Here it is for real: nickclegg.medium.com/you-and-the-al…
Aaaaaand here's the actual link. nickclegg.medium.com/you-and-the-al…
The most compelling pushback I’ve had on this is about losing long-term revenue if advertisers/users defect. Basically (1) FB is too dominant, there’s nowhere to defect TO and (2) businesses favor short-term profits over long-term all the time. 21/
I don’t think those negate the point. They complicate it in useful ways.

None of this can be explained by single causes or pithy explanations, including anything I said. (Not that I’m claiming pithiness in a thread this long.) 22/
The other point I see a lot — about a Clegg’s piece, not my thread — is that the whole framing (and title, good lord) lands like he’s saying “this is your fault, users!” I hadn’t seen that at first, but yeah. Comms fail.
And this on why “polarization” maybe only looks bad if you’re someone who benefits from the status quo. (To oversimplify—there’s a lot more to the thread.)

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Daphne Keller

Daphne Keller Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @daphnehk

Aug 28
I can't express how unutterably tired I feel after reading this absurd 3rd Cir ruling. It denies TikTok 230 immunity for a claim that is (very thinly) framed as liability based on algorithmic promotion, instead of liability based on user content. 1/

cases.justia.com/federal/appell…
This issue was fully briefed to the Supreme Court, with approximately infinity amicus briefs covering every possible angle, a little over a year ago. The Court decided not to decide. 2/

scotusblog.com/case-files/cas…
Now the 3rd Circuit is engaging in the absurd pretense that the Court actually decided this issue in Moody v. NetChoice, because it said algorithmic ranking that advances the platform's content moderation goals is the platform's 1st Am protected speech. Check out fn 13. 3/
Read 8 tweets
Aug 26
The Telegram CEO arrest in France seems unsurprising, and like something that also could have happened under U.S. law. 1/
It has long been rumored (and maybe reliably reported?) that Telegram fails to remove things like unencrypted CSAM or accounts of legally designated terrorist organizations even when notified.

That could make a platform liable in most legal systems, including ours.
2/
CSAM, terrorist content, and drug sales are all regulated by federal criminal law. Platforms have no immunity from that law. There are even some special provisions for platforms in federal criminal drug law (though IIRC they didn't add much). 3/
Read 6 tweets
Aug 19
I’ve been playing with a special “Trust and Safety regulation expert” version of ChatGPT. It is remarkable how much it thinks that US and EU law require platforms to suppress expression that is perfectly legal in those jurisdictions. 1/
As it turns out, regular ChatGPT gives very similar answers, telling platforms to remove "harmful" user speech. This isn’t AI going rogue or hallucinating. This is AI reflecting what the laws and secondary materials it trained on actually say. 2/
For the EU, I think the training materials actually led the AI to misstate the law, saying that “risk mitigation” means suppressing users ability to see and share lawful content. Here’s the thread (and link to Google doc with its answers) on that. 3/
Read 11 tweets
Sep 29, 2023
The S Ct will review the must-carry provisions of the TX and FL laws, and the requirements for "individualized notice" to users of content moderation decisions, but not other transparency requirements in the laws.

It says cert is for Questions 1 and 2 in the SG's brief. 1/
What statutory provisions does that actually encompass? The SG brief says it includes Texas's appeals provision, too. The Texas statutory sections it mentions as part of Q2 are 120.103 and 120.104, unless I am missing something. 2/
Here is my annotated copy of the Texas law, BTW. 3/
docs.google.com/document/d/1DP…
Read 15 tweets
Sep 26, 2023
The EU's database is live! In theory it should include every Statement or Reasons sent by platforms explaining content moderation decisions. I've groused about it, but it's an amazingly ambitious effort and already pretty interesting. 1/
When I first opened it half an hour or so ago, the database had 3.4 million entries. Now it's 3.5 million. 2/
Tiktok has submitted 1,764,373 Statements of Reason. X has submitted TWO.
You can hear the enforcers in Brussels salivating from all the way over here in California. 3/
Read 12 tweets
Jul 21, 2023
The statements from Thierry Breton of the European Commission about shutting down social media during riots are shocking. They vindicate every warning about the DSA that experts from the majority world (aka global south) have been shouting throughout this process. 1/
Breton asserts authority under the DSA to make platforms remove posts calling for “revolt” or “burning of cars” immediately. If platforms don’t comply, they will be sanctioned immediately and then banned. (He says “immediately” four times in a short quote.) 2/
Image
Image
As someone who generally defends the DSA as an instrument that (1) has a lot of process constraining such extreme exercise of state power and (2) will be enforced by moderate and rights-respecting regulators, Breton’s take on the DSA me feel like a naive chump. 3/
Read 15 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(