Surprising, But Important: Facebook Sorta Shuts Down Its Face Recognition System

from the good-to-see dept

A month ago, I highlighted how Facebook seemed uniquely bad attaking a long term view and publicly committing to doing things that are good for the world, but bad for Facebook in the short run . So it was a bit surprising earlier this week to see Facebook (no I'm not calling it Meta, stop it) announce that it was shutting down its Face Recognition system and (importantly) deleting over a billion "face prints" that it had stored.

The company's announcement on this was (surprisingly!) open about the various trade-offs here, both societally and for Facebook, though (somewhat amusingly) throughout the announcement Facebook repeatedly highlights the supposed societal benefits of its facial recognition.

Making this change required careful consideration, because we have seen a number of places where face recognition can be highly valued by people using platforms. For example, our award-winning automatic alt text system, that uses advanced AI to generate descriptions of images for people who are blind and visually impaired, uses the Face Recognition system to tell them when they or one of their friends is in an image.

[....]

But the many specific instances where facial recognition can be helpful need to be weighed against growing concerns about the use of this technology as a whole. There are many concerns about the place of facial recognition technology in society, and regulators are still in the process of providing a clear set of rules governing its use. Amid this ongoing uncertainty, we believe that limiting the use of facial recognition to a narrow set of use cases is appropriate.

One interesting tidbit buried in this is that only about 1/3 of Facebook users opted in to use Facebook's facial recognition tool (despite the company pushing it heavily on users). At the very least, it showed that a large number of users weren't comfortable with the technology.

There's also the issue that, while they're turning off the tool and deleting the facial prints, the NY Times notes they're hanging on to the algorithm that was built on all those faces:

Although Facebook plans to delete more than one billion facial recognition templates, which are digital scans of facial features, by December, it will not eliminate the software that powers the system, which is an advanced algorithm called DeepFace. The company has also not ruled out incorporating facial recognition technology into future products, Mr. Grosse said.

That's resulted in some (expected) amount of cynicism from Facebook's critics that Facebook "got what it wanted" and is now moving on. However, I think that's a bit silly. Facebook could have easily kept the facial recognition program going. Of all the regulatory pressures the company is facing, this was way down the list and barely on the radar.

And, to make a bigger point, here's a case where the company is actually doing the right thing: turning off a questionable product and deleting a ton of data it collected. And we should at least encourage both Facebook and other companies to be willing to make that decision based on recognizing the societal risks, and without waiting around until they're forced to do so.

Filed Under: ai, data, deepface, facial recognition, privacy, society, trade-offs
Companies: facebook

The Scale Of Content Moderation Is Unfathomable

from the it's-way-more-than-you-think dept

Sometimes it's difficult to get across to people "the scale" part when we talk about the impossibility of content moderation at scale. It's massive. And this is why whenever there's a content moderation decision that you dislike or that you disagree with, you have to realize that it's not personal. It wasn't done because someone doesn't like your politics. It wasn't done because of some crazy agenda. It was done because a combination of thousands of people around the globe and still sketchy artificial intelligence are making an insane number of decisions every day. And they just keep piling up and piling up and piling up.

Evelyn Douek recently gave a (virtual) talk at Stanford on The Administrative State of Content Moderation, which is worth watching in its entirety. However, right at the beginning of her talk, she presented some stats that highlight the scale of the decision making here. Based on publicly revealed transparency reports from these companies, in just the 30 minutes allotted for her talk, Facebook would take down 615,417 pieces of content, YouTube would take down 271,440 videos, channels, and comments, and TikTok would take down 18,870 videos. And, also, the Oversight Board would receive 48 petitions to review a Facebook takedown decision.

And, as she notes, that's only the take down decisions. It does not count the "leave up" decisions, which are also made quite frequently. Facebook is not targeting you personally. It is not Mark Zuckerberg sitting there saying "take this down." The company is taking down over a million pieces of content every freaking hour. It's going to make mistakes. And some of the decisions are ones that you're going to disagree with.

And, to put that in perspective, she notes that in its entire history, the US Supreme Court has decided a grand total of approximately 246 1st Amendment cases, or somewhere around one per year. And, of course, in those cases, it often involves years of debates, and arguments, and briefings, and multiple levels of appeals. And sometimes the Supreme Court still gets it totally wrong. Yet we expect Facebook -- making over a million decisions to take content down every hour -- to somehow magically get it all right?

Anyway, there's a lot more good stuff in the talk and I suggest you watch the whole thing to get a better understanding of the way content moderation actually works. It would be helpful for anyone who wants to opine on content moderation to not just understand what Douek is saying, but to really internalize it.

Filed Under: content moderation, due process, evelyn douek, scale, takedowns
Companies: facebook, tiktok, youtube

The Internet Is Not Facebook; Regulating It As If It Were Will Fuck Things Up

from the fixed-that-for-you dept

I've mocked the NY Times for its repeated failures to understand basic facts about internet regulations such as Section 230 -- but the organization also deserves credit when it gets things (mostly) right. Last week, Farhad Manjoo wrote up a great opinion piece noting that, even if you agree that Facebook is bad, most regulatory proposals would make things much, much worse.

He focuses on the blatantly unconstitutional "Health Misinformation Act" from Senators Klobuchar and Lujan, which would appoint a government official to declare what counts as health misinformation, and then remove Section 230 protections from any website that has such content. As Manjoo rightly notes, it's as if everyone has forgotten who was President from 2017 to early 2021 and hasn't considered what he or someone like him would do with such powers:

There’s only one problem: What is health misinformation? I know of no oracular source of truth about Covid-19. Scientific consensus has shifted dramatically during the pandemic, and even now experts are divided over important issues, such as whether everyone should get a vaccine booster shot. Klobuchar and Luján’s bill elides these complications. Instead they designate an all-knowing authority: Health misinformation, the bill says, is whatever the secretary of health and human services decides is health misinformation.

I’m sorry — what? Have the senators forgotten that just last year we had a president who ridiculed face masks and peddled ultraviolet light as a miracle cure for the virus? Why would we choose to empower such a president’s cabinet appointee as the arbiter of what’s true and false during a pandemic? And not just a pandemic — since the law defines a public health emergency so broadly, I wouldn’t put it past a science-averse future secretary from attempting to declare discussions about abortion, birth control, transgender health or whatever else as “misinformation.”

As he notes, so many of the proposals out there fill him "with deep dread" and they should. They're "this is bad, we must do something, this is something" proposals with little regard to (1) whether or not they are constitutional, and (2) whether or not they'd do anything to help with whatever the "bad" thing is. And, a key point, many of them would help Facebook at a moment when Facebook is losing users to competitors:

Rather than curbing the influence of Big Tech, altering Section 230 might only further cement Facebook and other tech giants’ hold over public discourse — because the giants might be the only companies with enough resources to operate under rules in which sites can be inundated with lawsuits over what their users post. Smaller sites with fewer resources, meanwhile, would effectively be encouraged to police users’ content with a heavy hand. It is no accident that Facebook has been telling lawmakers that it welcomes reforms to Section 230 — while smaller sites like Etsy and Tripadvisor are nervous about the possibility.

What's clear (though Manjoo doesn't say it in the piece) is that so many of these regulations are really targeted at Facebook, and don't realize how much they'd impact the wider ecosystem. They're regulating the internet as if Facebook was the internet. And the end result might be that... it becomes a self-fulfilling prophecy. If only Facebook can survive under these regulations, then Facebook becomes the internet.

There is one point in the article that I think is questionable, though. Manjoo spoke to NY Times suer (though he eventually dropped the suit) Larry Lessig, who told him:

Lawrence Lessig, a professor at Harvard Law School who has been working with Haugen, the Facebook whistle-blower, told me that some content-neutral rules for online speech might survive constitutional scrutiny — for example, a rule that set an upper limit on the number of times a Facebook post could be reshared.

And... I may not be a Harvard law professor, but that's just wrong. Limiting how often something can be shared is clearly a 1st Amendment issue, because it is the literal suppression of speech. It may be "content neutral" but that's not the only part of the strict scrutiny test used to judge whether or not something can get around the 1st Amendment. They also need to be narrowly tailored, be the least restrictive means to achieve their result and leave open ways for speech to be spread -- and I can't see how restricting how often something can be shared can possibly meet that bar.

Either way, kudos to Manjoo and the NY Times for highlighting that even if you think Facebook is awful, most regulatory proposals would probably make the internet, and Facebook, way worse.

Filed Under: internet regulations, section 230
Companies: facebook

Forget 'The Kids These Days'; It's The Adults And Their Moral Panics To Worry About

from the truly-devious-licks dept

A recent episode of the Reply All podcast, Absolutely Devious Lick, touched on a bunch of interesting points regarding the never-ending debates about social media, content moderation, and how it's supposedly damaging the kids these days. It's worth listening to the entire episode, but it begins by talking about a very slightly viral TikTok "challenge" which became known as Devious Licks -- lick being slang for something you stole. It started with a kid putting up a TikTok video of him holding a box of disposable masks, suggesting that he had stolen it from the school. Because school kids sometimes do stupid things to copy their stupid friends, a few others posted similar videos, including one early one of a kid taking a soap dispenser. And then there were some stories of it spreading and people going more extreme, because, you know, kids. But it didn't seem to spread that far initially.

But, of course, the thing became a lot more viral after mainstream media jumped on it with their typical "OMG, the kids these days" kind of coverage, starting with the New York Times, CNN, USA Today and then like every random local news jumping on the trend to tsk tsk about the kids these days.

Prominent grandstanding Senator, Richard Blumenthal called on TikTok execs to testify over all of this, which turned into another ridiculous Senate hearing in which old men yell at social media execs about how they're harming kids.

But, scratch the surface a little, and beyond a few dumb kids, this seems a lot more like adults over-reacting and freaking out, and making the story go much, much, much more viral than it did in reality. Indeed, the only news organization I've seen that recognized that most of this was a moral panic by adults was Curbed, which noted that, yes, there was some actual vandalism done by kids, but a lot of it seemed to be kids mocking the trend as well:

For all the real-life vandalism, what is also very real is that some of these teen punks might be punking us all too. One student, Gavino, a 17-year-old high schooler in Minnesota, uploaded a video showing a classroom sink gushing water, describing it as a “Devious Lick” gone wrong. When I contacted him over Snapchat, however, he told me he wasn’t actually trying to poach the faucet. It was broken, so he made a video about it. When I reached out to another student, a 14-year-old, who posted a TikTok stuffing Chromebooks into his backpack, he told me he didn’t even take them out of the building.

“Seeing people do it, others think they can ‘one-up’ the last person and get something better without getting caught,” Gavino said, adding, “It’s just being funny, trying to get five seconds of fame on a big platform.” Said a 17-year-old girl in L.A., who posted a video “stealing” a microscope from her science lab (she actually owned it, and she filmed the TikTok at home), “I made the video because obviously it was trending. Mostly, it’s for internet clout. And to be funny. It’s not a ‘fitting in’ type of thing. It’s literally just for clout, to show off or … whatever.”

Meanwhile, pretty damn early in all of this, TikTok banned the "Devious Lick" tag and told people searching for it to knock it off:

Also, as far as I can tell, none of the media orgs that covered the whole moral panic freak-out noted that a bunch of kids started to counterprogram whatever vandalism occurred in the opposite direction, posting the opposite of "devious licks": angelic yields, where they would show themselves adding new items to schools (often school bathrooms, since so many of the devious licks stories were about soap and toilet paper being taken from bathrooms).

And that brings us back around to the Reply All episode, which followed a secondary freak-out, after the Devious Licks challenge, in which a document was being passed around claiming to have pre-planned a bunch of other "challenges" for kids in school throughout the rest of the school year. Each month, the posting suggested, kids were planning to do crazy stupid shit (I mean, more crazy and more stupid than any normal teenager) in schools... for TikTok. But, as Reply All's Anna Foley noted, everything about the story just seemed weird -- including (1) teenagers actually planning shit out for an entire school year, and (2) the language on the document didn't sound at all like kids ("slap a teacher on the backside"?!?!?).

Foley started investigating and... basically tracked it down to adults freaking out. She found a School Resource Officer (SRO) who had been early posting the list to Facebook, who didn't seem particularly concerned at all whether or not it was accurate or not -- taking the typical "better safe than sorry" kind of approach. And from there, she traced it back to a school superintendent who claimed she had gotten it from students, though wouldn't say who. But what becomes pretty clear is that almost no kids were passing it around or seriously considering it.

Indeed, there's a hilarious moment in the podcast in which Foley quotes posts from kids laughing and mocking all the adults falling for this nonsense.

It's hard not to look at this like any other moral panic by adults who somehow have forgotten what it's like to be a teenager. Are there teenagers doing stupid stuff -- including some theft and vandalism? Yes, of course. Because that's something teenagers do. I mean, I did stupid shit as a teenager too. But, the idea that this is something new, or is caused by social media doesn't really get much support in reality. It does seem like the real "disinformation" came from the adults, not the kids, and the people who bought into it were the adults, not the kids -- and the leading vector of it being sent around was the mainstream media, not social media.

But I don't see Senator Blumenthal calling the head of CNN, the NY Times, and USA Today to come testify about their role in spreading "devious lick" news to gullible adults. Because then he'd have to admit that he, too, is a silly gullible adult.

Filed Under: adults, devious licks, kids, moral panic, pranks, social media
Companies: facebook, tiktok

It's Ridiculous The 'Developing World' Wasn't Given Access To The Facebook Files

from the do-this-the-right-way dept

By now it's fairly clear the Facebook leaks showcase a company that prioritized near-mindless international growth over the warnings of their own experts. They also show a company that continues to painfully struggle to be even marginally competent at scale, whether we're talking about content moderation or rudimentary customer service. While this has become an all-encompassing media spectacle, the real underlying story isn't particularly unique. It's just a "growth for growth's sake" mindset, where profit and expansion trumped all reason. It just happens to be online, and at unprecedented international scale.

One thing I haven't seen talked a lot about is the fact that if you look back a few years, an awful lot of of folks in developing nations saw these problems coming a mile away, long before their Western counterparts. For a decade, international activists warned repeatedly about the perils of Facebook's total failure to understand the culture/regulations/language/norms of the countries they rapidly flooded into. Yet bizarrely, Frances Haugen's PR team somehow excluded most of these countries when it came time to recently release access to the Facebook files:

The "consortium" handling release of the files is basically a loose collaboration between 17 U.S. news orgs and a handful of European outlets. They're all being given access to redacted versions of documents Haugen provided the Securities and Exchange Commission, showing Facebook repeatedly prioritized growth and profit over, well, everything else. The whole thing was handled by Haugen's PR team via an ordinary embargo, which some oddly saw as itself somehow nefarious (it's not, embargoes, though often kind of stupid, are commonly used to maximize impact).

The real problem was who was included in that consortium. And a bigger problem, oddly not talked about until a month into the Facebook leak news cycle, was that much of the developed world was just... excluded... from the coalition by her PR reps. The exclusion of academics and researchers that could make the most sense of the data was a problem. But restricting analysis to most white, western newsrooms, (despite Haugen's very clear understanding that most of Facebook's impact problems disproportionately harmed developing nations) is particularly odd and tasteless.

For all the problems Facebook (sorry, Meta) has had in the United States in regards to managing the company's platform at scale, those problems have been dramatically worse internationally. Facebook was so excited to flood into dozens of international locations to wow investors, they didn't dedicate the time, resources, or attention needed to actually understand what they were doing. Even if they had (and the whistleblowers keep showing they absolutely didn't), the sheer scale of the expansion made it impossible to do well. That Facebook did so anyway despite being warned about it is an act of greed and hubris.

It was actually a net neutrality debate that keyed many overseas activists into Facebook's problems more than a decade ago. Activists in India were particularly sensitive to Facebook's attempts to conflate "the internet" with Facebook in developing nations. If you recall, activists in India successfully derailed Facebook's Free Basics program, which was Facebook's attempt to corner developing nation ad markets under the banner of altruism.

Basically, it involved Facebook striking deals with local wireless companies to offer discounted access to Facebook, under claims that aggressively curated version of "online" was better than no online access at all. It was a highly curated walled garden bastardized variation of AOL or CompuServe, in which Facebook decided what information, services, and resources mattered (they initially even banned encrypted services). But activists and international experts were quick to see the problem with giving Facebook this kind of power, especially in countries they didn't take the time to understand.

We've seen repeatedly how conflating "Facebook" with "the internet" (or Whatsapp with "the internet") has created a laundry list of problems that are especially pronounced in developing nations. The centralized approach of programs like Free Basics defeated the purpose of the open internet, reduced transparency, was a big boon for authoritarian governments, and helped create an even more potent funnel for propaganda. One recurring theme in whistleblower accounts is that Facebook's own researchers generally warned about all of this, repeatedly, but were ignored for profit and growth's sake.

As early as 2015 organizations like Mozilla were busy arguing that if Facebook genuinely cared about information access in developing nations, they should simply fund access to the internet itself. Facebook was ultimately forced to back off its plan in some countries like India and Egypt, but if you were a reporter or activist in these countries who pointed out the problems with Facebook's ambition, you were accused of being an enemy of the poor. When the Free Basics brand became toxic, Facebook just named Free Basics something else (sound familiar?).

There has been some valid and not so valid criticism of the way Haugen handled these latest revelations. Some have tried to argue that because she was smart enough to hire lawyers and a PR team to maximize impact she can't possibly be technically seen as a "whistleblower." There was also some brief hyperventilation over a Politico report, with some trying to claim that because she had received some money from investor Pierre Omidyar, she shouldn't be taken seriously. But a shrewd, organized whistleblower is still a whistleblower, and Omidyar proxy groups actually just funded whistleblower orgs Haugen was part of after she went public. It's actually a good thing to see a whistleblower do the right thing and not be economically and reputationally devastated for once.

But it's both weird and telling that people freaked out about these perceived injustices, but didn't notice that the whistleblower's PR coalition apparently just forgot the developing world existed:

Nobody I've talked to so far at news organizations seems clear why this seems to have happened (a strange decision for an effort geared toward greater transparency). It's not like it would be particularly difficult to coordinate the release via the same organizations in places like India that warned about Facebook's consolidated power almost a decade ago (see: IFEX). Some outlets, like Gizmodo, have been trying to expand access to the source documents to everyone. That's apparently to the chagrin of Haugen's PR team, who seems to think they can put the genie back in the bottle.

There was a certain hubris in Facebook stumbling its way across the developing world in a quest for growth without bothering to understand the impact their platform would have on foreign cultures. But there's a fairly substantial amount of hubris in excluding these developing nations from accessing raw data on a problem they've disproportionally been harmed by.

Filed Under: bias, developing nations, facebook files, facebook papers, frances haugen, india, journalists
Companies: facebook, meta

Let Me Rewrite That For You: Washington Post Misinforms You About How Facebook Weighted Emoji Reactions

from the let's-clean-it-up dept

Journalist Dan Froomkin, who is one of the most insightful commentators on the state of the media today, recently began a new effort, which he calls "let me rewrite that for you," in which he takes a piece of journalism that he believes misled readers, and rewrites parts of them -- mainly the headline and the lede -- to better present the story. I think it's a brilliant and useful form of media criticism that I figured I might experiment with as well -- and I'm going to start it out with a recent Washington Post piece, one of many the Post has written about the leaked Facebook Files from whistleblower Frances Haugen.

The piece is written by reporters Jeremy Merrill and Will Oremus -- and I'm assuming that, like many mainstream news orgs, editors write the headlines and subheads, rather than the reporters. I don't know Merrill, but I will note that I find Oremus to be one of the most astute and thoughtful journalists out there today, and not one prone to fall into some of the usual traps that journalists fall for -- so this one surprised me a bit (though, I'm also using this format on an Oremus piece, because I'm pretty sure he'll take the criticism in the spirit intended -- to push for better overall journalism on these kinds of topics). The article's headline tells a story in and of itself: Five points for anger, one for a ‘like’: How Facebook’s formula fostered rage and misinformation, with a subhead that implies something similar: "Facebook engineers gave extra value to emoji reactions, including ‘angry,’ pushing more emotional and provocative content into users’ news feeds." There's also a graphic that reinforces this suggested point: Facebook weighted "anger" much more than happy reactions. And it's all under the "Facebook under fire" designation:

Seeing this headline and image, it would be pretty normal for you to assume the pretty clear implication: people reacting happily (e.g. with "likes") on Facebook had those shows of emotions weighted at 1/5th the intensity of people reacting angrily (e.g. with "anger" emojis) and that is obviously why Facebook stokes tremendous anger, hatred and divisiveness (as the story goes).

But... that's not actually what the details show. The actual details show that initially when Facebook introduced its list of five different "emoji" reactions (to be added to the long iconic "like" button), it weighted all five of them as five times as impactful as a like. That means that "love," "haha," "wow," and "sad" also were weighted at 5 times a single like, and identical to "angry." And while the article does mention this in the first paragraph, it immediately pivots to focus only on the "angry" weighting and what that means. When combined with the headline and the rest of the article, it's entirely possible to read the article and not even realize that "love," "sad," "haha," and "wow" were also ranked at 5x a single "like" and to believe that Facebook deliberately chose to ramp up promotion of "anger" inducing content. It's not only possible, it's quite likely. Hell, it's how I read the article the first time through, completely missing the fact that it applied to the other emojis as well.

The article also completely buries how quickly Facebook realized this was an issue and adjusted the policy. While it does mention it, it's very much buried late in the story, as are some other relevant facts that paint the entire story in a very different light than the way many people are reading it.

As some people highlighted this, Oremus pointed out that the bigger story here is "how arbitrary initial decisions, set by humans for business reasons, become reified as the status quo." And he's right. That is the more interesting story and one worth exploring. But that's not how this article is presented at all! And, his own article suggested the "reified as the status quo" part is inaccurate as well, though, again, that's buried further down in the story. The article is very much written in a way where the takeaway for most people is going to be "Facebook highly ranks posts that made you angry, because stoking divisiveness was good for business, and that's still true today." Except none of that is accurate.

So... let's rewrite that, and try to better get across the point that Oremus claims was the intended point of the story.

The original title, again is:

Five points for anger, one for a ‘like’: How Facebook’s formula fostered rage and misinformation

Let's rewrite that:

Facebook weighted new emojis much more than likes, leading to unintended consequences

Then there's the opening of the piece, which does mention very quickly that it applied to all five new emojis, but quickly pivots to just focusing on the anger:

Five years ago, Facebook gave its users five new ways to react to a post in their news feed beyond the iconic “like” thumbs-up: “love,” “haha,” “wow,” “sad” and “angry.”

Behind the scenes, Facebook programmed the algorithm that decides what people see in their news feeds to use the reaction emoji as signals to push more emotional and provocative content — including content likely to make them angry. Starting in 2017, Facebook’s ranking algorithm treated emoji reactions as five times more valuable than “likes,” internal documents reveal. The theory was simple: Posts that prompted lots of reaction emoji tended to keep users more engaged, and keeping users engaged was the key to Facebook’s business.

Facebook’s own researchers were quick to suspect a critical flaw. Favoring “controversial” posts — including those that make users angry — could open “the door to more spam/abuse/clickbait inadvertently,” a staffer, whose name was redacted, wrote in one of the internal documents. A colleague responded, “It’s possible.”

The warning proved prescient. The company’s data scientists confirmed in 2019 that posts that sparked angry reaction emoji were disproportionately likely to include misinformation, toxicity and low-quality news.

Let's rewrite that, both using what Oremus claims was the "bigger story" in the article, and some of the information that is buried much later.

Five years ago, Facebook expanded the ways that users could react to posts beyond the iconic "like" thumbs-up, adding five more emojis: "love," "haha," "wow," "sad," and "angry." With this new addition, Facebook engineers needed to determine how to weight these new engagement signals. Given the stronger emotions portrayed in these emojis, the engineers made a decision that had a large impact on how the use of those emojis would be weighted in determining how to rank a story: each of those reactions would count for five times the weight of the classic "like" button. While Facebook did publicly say at the time that the new emojis would be weighted "a little more" than likes, and that all the new emojis would be weighted equally, it did not reveal that the weighting was actually five times as much.

This move came around the same time as Facebook's publicly announced plans to move away from promoting clickbait-style news to users, and to try to focus more on engagement with content posted by friends and family. However, it turned out that friends and family don't always post the most trustworthy information, and by overweighting the "emotional" reactions, this new move by Facebook often ended up putting the most emotionally charged content in front of users. Some of that content was joyful -- people reacting with "love" to engagements and births -- but some of it was disruptive and divisive, people reacting with "anger" to false or misleading content.

Facebook struggled internally with this result -- while also raising important points about how "anger" as a "core human emotion" is not always tied to bad things, and could be important for giving rise to protest movements against autocratic and corrupt governments. However, since other signals were weighted significantly more than even these emojis -- for example, replies to posts had a weight up to 30 times a single "like" click -- not much was initially done to respond to the concerns about how the weighting on anger might impact the kinds of content users were prone to see.

However, one year after launch, in 2018, Facebook realized weighting "anger" so highly was a problem, and downgraded the weighting on the "anger" emoji to four times a "like" while keeping the four other emoji, including "love," "wow," and "haha" at five times a like. A year later, the company realized this was not enough and even though "anger" is the least used emoji, by 2019 the company had put in place a mechanism to "demote" content that was receiving a disproportionate level of "anger" reactions. There were also internal debates about reranking all of the emoji reactions to create better news feeds, though there was not widespread agreement within the company about how best to do this. Eventually, in 2020, following more internal research on the impact of this weighting, Facebook reweighted all of the emoji. By the end of 2020 it had cut the weight of the "anger" emoji to zero -- taking it out of the equation entirely. The "haha" and "wow" emojis were weighted to one and a half times a like, and the "love" and "sad" were weighted to two likes.

From there, the article could then discuss a lot of what other parts of the article does discuss, about some of the internal debates and research, and also the point that Oremus raised separately, about the somewhat arbitrary nature of some of these ranking systems. But I'd argue that my rewrite presents a much more accurate and honest portrayal of the information than the current Washington Post article.

Anyone know how I can send the Washington Post an invoice?

Filed Under: algorithm, emoji, facebook papers, framing, frances haugen, journalism, let me rewrite that for you, ranking, reactions
Companies: facebook

When Facebook Turned Off Its News Feed Algorithm, It Made Everyone's Experience Worse... But Made Facebook More Money

from the oh,-look-at-that dept

For reasons I don't fully understand, over the last few months, many critics of "big tech" and Facebook, in particular, have latched onto the idea that "the algorithm" is the problem. It's been almost weird how frequently people insist to me that if only social media got rid of algorithmically recommending stuff, and went back to the old fashioned chronological news feed order, all would be good in the world again. Some of this seems based on the idea that algorithms are primed to lead people down a garden path from one type of video to ever more extreme videos (which certainly has happened, though how often is never made clear). Some of it seems to be a bit of a kneejerk reaction to simply disliking the fact that these companies (which many people don't really trust) are making decisions about what you may and may not like -- and that feels kinda creepy.

In the past few weeks, there's been a bit of a fever pitch on this topic, partly in response to whistleblower Frances Haugen's leak of documents, in which she argues that Facebook's algorithm is a big part of the problem. And then there's the recent attempt by some Democrats in Congress to take away Section 230 from algorithmically recommended information. As I noted, the bill is so problematic that it's not clear what it's actually solving.

But underlying all of this is a general opinion that "algorithms" and "algorithmic recommendations" are inherently bad and problematic. And, frankly, I'm confused by this. At a personal level, the tools I've used that do algorithmic recommendations (mainly: Google News, Twitter, and YouTube) have been... really, really useful? And also pretty accurate over time in learning what I want, and thus providing me more useful content in a more efficient manner, which has been pretty good for me, personally. I recognize that not everyone has that experience, but at the very least, before we unilaterally declare algorithms and recommendation engines as bad, it might help to understand how often they're recommending stuff that's useful and helpful, as compared to how often they're causing problems.

And, for all the talk about how Haugen's leaking has shown a light on the "dangers" of algorithms, the actual documents that she's leaked might suggest something else entirely. Reporter Alex Kantrowitz has reported on one of the leaked documents, regarding a study Facebook did on what happens when Facebook turns off the algorithmic rankings and... it was not pretty. But, contrary to common belief, Facebook actually made more money without the News Feed algorithm.

In February 2018, a Facebook researcher all but shut off the News Feed ranking algorithm for .05% of Facebook users. “What happens if we delete ranked News Feed?” they asked in an internal report summing up the experiment. Their findings: Without a News Feed algorithm, engagement on Facebook drops significantly, people hide 50% more posts, content from Facebook Groups rises to the top, and — surprisingly — Facebook makes even more money from users scrolling through the News Feed.

Considering how often we've heard, including from Haugen herself, that Facebook's decision-making is almost always driven by what will beneficially impact the bottom line the most, this deserves some consideration. Because the document... suggests something quite different. In fact, what the researchers seemed to find was that people hated it, but it made them spend more time on the site and see more ads because they had to poke around to try to find the interesting stuff they wanted to see, and that drove up ad rates. If Facebook were truly focused on just the bottom line, then, they should consider turning off the news feed algorithm -- or, just supporting the awful JAMA bill in Congress which will create incentives for the same result:

Turning off the News Feed ranking algorithm, the researcher found, led to a worse experience almost across the board. People spent more time scrolling through the News Feed searching for interesting stuff, and saw more advertisements as they went (hence the revenue spike). They hid 50% more posts, indicating they weren’t thrilled with what they were seeing. They saw more Groups content, because Groups is one of the few places on Facebook that remains vibrant. And they saw double the amount of posts from public pages they don’t follow, often because friends commented on those pages. “We reduce the distribution of these posts massively as they seem to be a constant quality compliant,” the researcher said of the public pages.

As always, there are lots of factors that go into this, and one experiment may not be enough to tell us much. Also, it's entirely possible that over time, the long term result would be less revenue because the increasing annoyances of not finding the more interesting stuff causes people to leave the platform entirely. But, at the very least, this leaked research pokes a pretty big hole in the idea that getting rid of algorithmic recommendations does anything particularly useful.

Filed Under: algorithms, chronological, facebook files, facebook papers, frances haugen, jama, leaks, news feed, section 230, user experience
Companies: facebook

Telecom, Broadcasters Convince FCC To Explore New Taxes On 'Big Tech'

from the this-probably-won't-go-well dept

Earlier this year, we noted how FCC Commissioner Brendan Carr had launched a bad faith effort suggesting that "big tech" gets a "free ride" on the internet, and should be forced to fund broadband expansion. Carr's argument, that companies like Google and Netflix somehow get a free ride (they don't) and should "pay their fair share," is a fifteen year old AT&T lobbyist talking point. AT&T's goal has always been to "double dip"; as in not only get paid for bandwidth by consumers and businesses, but to get an additional troll toll simply for, well, existing.

AT&T has long tried to offload its (often neglected or half-completed) network build and maintenance costs to somebody else to make investors happy. That somebody else is usually taxpayers, who've thrown billions in pointless tax breaks and dubious regulatory favors at the company in exchange for fiber networks that are always (so mysteriously!) left half completed and jobs that never arrive. Now AT&T (and their broadcaster allies) want tech giants to pay as well.

Over at the right wing Washington Examiner, you can see how this effort is framed in order to sell it to the public and regulators:

"The Federal Communications Commission is looking to hit Big Tech companies with new regulatory fees related to their high use of broadband facilitated by the agency — a sign of Washington's growing skepticism of Silicon Valley wealth and power."

You're given no indication that this is a telecom (AT&T) or broadcasting (the National Association of Broadcasters) bad faith political play. Carr (read: AT&T's policy folks) had originally suggested in a carefully seeded editorial that tech companies should be forced to pay into the Universal Service Fund, which helps fund broadband expansion to schools and low-income communities. Now, Carr's argument has also been piggybacked onto by broadcasters, who want tech companies to pay a tax simply to use unlicensed spectrum (like Wi-Fi hardware). Fortunately, the Examiner at least includes experts who point out the idea is stupid:

"Unlicensed spectrum was first established by the commission in 1985, and allows the general public to freely use, without a license, services such as Wi-Fi networks, Bluetooth waves, garage door openers, and other wireless technologies.

“The FCC is asking the public if we think fees on unlicensed spectrum is a good idea or not,” said Harold Feld, a telecom policy expert and lawyer at the consumer advocacy group Public Knowledge. “And I’m here to say it’s a bad idea. In the spirit of Halloween, I plan to take a chainsaw to it."

So I think the FCC is considering this route because it's true the USF needs more money if we're going to help make sure low-income kids and schools have decent connectivity. The problem is that the FCC's proposal is written in such a way that it sounds like you could potentially be seeing annoying new taxes on Wi-Fi networks, Bluetooth devices, garage door openers, and more. The other problem is the core motivation here is by telecom and broadcast companies who simply want to offload their costs to somebody else. If they're the ones writing the proposal (and it sounds like they are), the end product isn't going to be particularly balanced or productive.

But to me the bigger problem is that if we're genuinely interested in shoring up broadband funding and expanding access, we should first be targeting the billions upon billions in tax breaks, regulator favors, merger approvals, and other perks we give to telecom and broadcast giants (in Comcast and AT&T's case one in the same) in exchange for jack shit. The idea that we waste billions by throwing it mindlessly at regional monopolies seems like a good place to start if you're serious about reform. But please take a moment to notice that this idea never gets mentioned by folks like Carr. That's fairly telling.

Nothing gets fixed in U.S. broadband (especially affordability and access issues) without addressing regional monopolization and the state and federal corruption that protects it. But not only do we not do that, if you watch regulators and lawmakers on both sides of the aisle, they can barely even candidly acknowledge it's even a problem. As a baseline (and it has been the baseline for 30 straight years) that's not a great recipe for success.

Filed Under: big tech, brendan carr, broadband, broadcasters, fcc, free ride, net neutrality, subsidies, taxes, telcos, universal service fund
Companies: at&t, facebook, google

The Whistleblower And Encryption: Everyone Has An Angle, And Not Everyone Is A Policy Expert

from the nuance,-nuance,-nuance dept

Over the weekend, the Telegraph (not the most trustworthy or reliable in a batch of UK news organizations that have long had issues with accuracy in reporting) claimed that the latest (and most high profile) Facebook whistleblower, Frances Haugen, was prepared to come out against encryption. This (quite rightly) raised the hackles of multiple encryption experts. As people were getting pretty worked up about it, the Telegraph (silently, and without notice) changed the headline of the piece (from "Facebook whistleblower warns ‘dangerous’ encryption will aid espionage by hostile nations" to "Facebook whistleblower warns company's encryption will aid espionage by hostile nations") as well as the actual text of the story, to suggest a slightly more nuanced (but still not great) view -- effectively saying she supported encryption, but was concerned that Facebook would use encryption as a "see no evil" kind of blindfold to problems on its platform.

Ms Haugen said that she is generally pro-encryption, which enhances users’ privacy. However, she added that Facebook’s plan was also way for the company to “sidestep” harmful content happening on its platform rather than address it.

She said: “End-to-end encryption definitely lets them sidestep and go ‘look we can’t see it, not our problem’.”

Of course, context and motives matter here, and the Telegraph -- which tends to be quite supportive of the current UK government, seemed to be twisting Haugen's (admittedly confused) statement in support of UK Home Secretary Priti Patel's positively dangerous plan to get rid of end-to-end encryption in the UK. It sure looks like the Telegraph went looking for a way to support that argument, and used Haugen's words to that effect.

A few hours later, Haugen actually testified before a UK Parliamentary committee and claimed her words were taken out of context. She said that she's strongly pro-encryption... but then tried to claim that her comments to the Telegraph were more about how she doesn't trust Facebook to actually implement encryption. Which is... a strange and almost nonsensical claim.

“I want to be very, very clear. I was mischaracterised in the Telegraph yesterday on my opinions around end-to-end encryption,” she said. “I am a strong supporter of access to open source end to end encryption software.

“I support access to end-to-end encryption and I use open source end-to-end encryption every day. My social support network is currently on an open source end-to-end encryption service.”

[....]

“Facebook’s plan for end-to-end encryption — I think — is concerning because we have no idea what they’re doing to do. We don’t know what it means, we don’t if people’s privacy is actually protected. It’s super nuanced and it’s also a different context. On the open source end-to-end encryption product that I like to use there is no directory where you can find 14 year olds, there is no directory where you can go and find the Uighur community in Bangkok. On Facebook it is trivially easy to access vulnerable populations and there are national state actors that are doing this.

“So I want to be clear, I am not against end-to-end encryption in Messenger but I do believe the public has a right to know what does that even mean? Are they really going to produce end-to-end encryption? Because if they say they’re doing end-to-end encryption and they don’t really do that people’s lives are in danger. And I personally don’t trust Facebook currently to tell the truth… I am concerned about them misconstruing the product that they’ve built — and they need regulatory oversight for that.”

But... here's the thing: Haugen may be a wonderful data scientist. And, she may have done the world tremendous good by leaking tons of internal Facebook documents, giving the world some insight into what's going on at the company. But that doesn't make her an expert on encryption. And, it shows. As Alec Muffett, a security expert who actually used to work on encryption at Facebook, noted in a detailed thread, what Haugen is asking for here is dangerous and shows a real lack of understanding about encryption.

First, she claims that there should be a government review of any Facebook end-to-end encryption to make sure it's legit. And, yes, there are many reasons to not trust Facebook, but introducing the idea that government needs to review and approve encryption is worse. Is she completely unaware of the government's history of constantly trying to undermine and backdoor encryption? I mean, it's not exactly secret. And the US government has been trying to undermine and backdoor encryption pretty aggressively lately. Suggesting that there needs to be some government entity blessing the encryption opens the door to all sorts of mischief.

The separate issue is claiming that end-to-end encryption for Facebook is somehow different because you can use Facebook for more than just messaging, and it's bolted on to other services. Again, as Muffett explains, this kind of thinking is dangerous as well. It suggests that encrypted chat needs to be silo'd and kept distant from tons of internet services, when the reality is often that many more internet services should be embracing encryption much more widely to protect their users.

This is also why it's difficult to understand Haugen's claims -- as they seem somewhat contradictory. Even if we take the Telegraph's mission-driven editing with a grain of salt, Haugen doesn't deny her claim that encryption makes it harder to protect Uighurs:

“A key part of [Chinese operatives’] strategy was to send malware to Uighurs who lived in places that weren’t China, as if they could compromise one phone they could compromise a whole community. We said we won’t be able to see the malware anymore [with encryption].”

But, that's backwards. Do we think Uighurs will be more protected with encryption, or without it? As Riana Pfefferkorn pointed out just last week, encryption and security go hand in hand. It is not -- as law enforcement would falsely have you believe -- that encryption and security are at odds. Encryption provides security -- especially against oppressive governments trying to genocide and entire culture. Uighurs need encryption much more than they "need" Facebook to be able to see what the Chinese are doing to protect the Uighurs.

Haugen's statement on the Uighurs seems ridiculous when thought about: it's basically arguing that without encryption Facebook can better protect the Uighurs from the Chinese government. Does anyone actually believe that? Or would they be better off with access to encryption? They shouldn't necessarily rely on Facebook's encryption, but arguing that it shouldn't be there to better protect them is just silly.

Again, Haugen has likely done the world a great benefit in leaking a bunch of internal documents (I'll have more on those soon). But it's important to remember that just because she blew the whistle regarding Facebook research, it doesn't make her an expert on everything else. She's not an expert on content moderation, or antitrust, or encryption. She may be a useful source for exploring what Facebook's research showed, or some of Facebook's decision making, but it's depressing how quickly eager politicians looking to gain support for their already existing plans are exploiting her to argue for their position on topics she's really not qualified to comment on. Indeed, it's also dismissing the hard work of tons of actual experts on these topics, from practitioners in the field to the academics who study these issues.

Filed Under: encryption, frances haugen, going dark, tech policy, uk
Companies: facebook