Officially, Google’s preferred description for Project Dragonfly is “exploratory.” “This is an exploratory project and no decision has been made about whether we could or would launch,” the company said today, in the wake of a significant new report on its plans for a censored search engine in China.
”Exploratory” is a word that conjures the earliest stages of development — of the conquistador, standing at the shore of a new land, surveying rough its shape and character. But according to Thursday’s report by The Intercept’s ace Dragonfly chronicler Ryan Gallagher, Google’s explorations had already reached the final stages of development by the time word of the project leaked.
And as Gallagher describes it, that was the result of a concerted effort by one of the project’s leaders designed to ensure that no word of the project leak — that the explorers would have already colonized the new land before anyone else had realized they had even set sail:
Yonatan Zunger, then a 14-year veteran of Google and one of the leading engineers at the company, was among a small group who had been asked to work on Dragonfly. He was present at some of the early meetings and said he pointed out to executives managing the project that Chinese people could be at risk of interrogation or detention if they were found to have used Google to seek out information banned by the government.
Scott Beaumont, Google’s head of operations in China and one of the key architects of Dragonfly, did not view Zunger’s concerns as significant enough to merit a change of course, according to four people who worked on the project. Beaumont and other executives then shut out members of the company’s security and privacy team from key meetings about the search engine, the four people said, and tried to sideline a privacy review of the plan that sought to address potential human rights abuses.
Zunger is the first person who worked on Dragonfly to speak about it publicly. (He left Google last year.) He and other sources tell Gallagher that Beaumont tried to bring Dragonfly into existence by bypassing key oversight measures, such as a privacy review. “His ideal circumstance was that most people would find out about this project the day it launched,” one source told Gallagher.
On one hand, this is understandable. Technically, the fact that China censors search results is a state secret. But for the Googlers who have been asked to work on the project, that has come as cold comfort.
Google says “privacy reviews are non-negotiable and we never short-circuit the process.” But current and former Googlers have been riled by the revelations in Gallagher’s report. Brandon Downey, who worked on projects related to Google’s 2006 move into China, called Thursday’s “the most jaw dropping of the Dragonfly stories.” The report suggests Google CEO Sundar Pichai lied to the public and to employees, Downey said. (Pichai has repeatedly characterized the project as exploratory.)
Meanwhile Liz Fong-Jones, a vocal internal critic who has pledged to quit Google in February if the company does not make significant policy changes, may be organizing a strike. In a Twitter thread, she asked fellow coworkers to put money into a strike fund that would help cover employees’ expenses during an extended walkout. Within hours, employees had raised $100,000.
The outpouring of funds reflects both Googlers’ growing sense of their own power, after the success of the walkout earlier this month, and their disdain of Project Dragonfly. Not everyone at the company shares that feeling. Just before publish time today, Google’s director of security and privacy, Heather Adkins, said she had not been sidelined during Dragonfly’s development.
But enough employees feel strongly about it that they are now organizing a labor action over it, in public, on Twitter. This is an extraordinary turn of events. At CNBC, Jillian D’Onfro talked to another of the contributors:
Current employee Irene Knapp, who previously made a public statement criticizing Google’s diversity efforts at Alphabet’s shareholder meeting earlier this year, also contributed to the fund.
“I’m proud to support this fund, it’s a great step forward,” Knapp says. “Everyone participating in labor efforts faces personal risks, and everyone’s risks are different. Those of us in a position to make it easier for others have an obligation to do so.”
What would trigger a strike? Fong-Jones suggested that Google would have to cross a red line — launching Dragonfly in China without a proper review of Dragonfly’s privacy implications, for example. “I firmly suggest that my current fellow colleagues think about what they’d do if the red line were crossed and an executive overrode a S&P launch bit, or members of the S&P team indicated that they were coerced into marking it green,” Fong-Jones wrote, referring to the company’s security and privacy teams.
How far away are we from that red line? Gallagher reports that at one point, the Dragonfly team was told to prepare to launch between January and April of next year. Given the current controversy — and the ongoing US trade war with China — that timeline almost certainly has been pushed back. In the meantime, the issue that will define Sundar Pichai’s legacy as CEO has taken a dark new turn. And as of today, his internal opposition is vocal, public, and funded.
Facebook released more details on its forthcoming disclosure requirements for UK political advertisers. As in other countries where the requirements have been introduced, UK advertisers will have to verify their identities before they can post political advertisements. The policy will include an exemption for news sites, Alex Hern reports, and that exemption could eventually come to the United States and Brazil, where publishers have also requested it.
Since the company first required news providers to register, Facebook’s Rob Leathern said, it had “built more controls to help prevent politically motivated actors looking to use false news or sensationalism as weapons, and in September, we announced a news indexing process designed to more clearly and consistently identify pages posting news on Facebook”.
Amazon is facing a new probe related to its status as both a retailer and a storefront for other storefronts, Sara Germano reports:
Germany’s antitrust regulator is opening an investigation into AMazon, making it the latest authority in Europe to examine the tech behemoth’s practices amid growing concern that the retailer might be unfairly leveraging its dominant position.
The Federal Cartel Office said Thursday it was examining business practices and terms on Amazon’s German marketplace—a part of the Amazon website that is open to third-party sellers. The probe follows “many complaints” from sellers on the online marketplace, the largest in Germany, according to the watchdog’s president, Andreas Mundt.
Facebook considered charging third parties for access to user data several years ago, Deepa Seetharaman and Kirsten Grind report, citing company emails. Lots of people yelled about this on Twitter yesterday, but my own view is that smart companies tend to consider everything when it comes to revenue.
Fallout from the Definers scandal is spreading to its other clients, including the UK’s Conservative party, Jamie Doward reports:
The Conservative party is under pressure to reveal details about its relationship with the London arm of a US lobbying firm accused of smear tactics against critics of Facebook.
UK Policy Group, a consultancy with close links to the Conservative party, is part of Definers Public Affairs, the controversial firm ditched by Facebook earlier this month following a New York Times exposé that has further dented the social media network’s image.
The military plans to purchase as many as 100,000 HoloLens augmented reality devices, Joshua Brustein reports. As a result, I have revised my projected HoloLens sales upward, to 100,004.
“Augmented reality technology will provide troops with more and better information to make decisions. This new work extends our longstanding, trusted relationship with the Department of Defense to this new area,” a Microsoft spokesman said in an emailed statement.
The U.S. Army and the Israeli military have already used Microsoft’s HoloLens devices in training, but plans for live combat would be a significant step forward.
Teens, for one, would like to tell you that they enjoy their social media apps, according to new data from Pew Research. Here’s Katie Notopoulos:
The vast majority of teens in a new survey believe social media is actually good for them, a Pew Research Center report shows. Seriously: 81% of teens said it makes them feel more connected to friends, 71% said it helps them show their creative side, 69% said it helps them make friends and with a more diverse group of people, and 68% feel like they have people who support them through tough times. (And no, this wasn’t secretly funded by Facebook.)
Much research has focused on social media being a huge waste of time at best, a facilitator of ideological bubbles, and a dangerous, hostile experience for young people at worst. But the 743 teens Pew surveyed say it’s actually, well, good. Millennials were the first to make social media mainstream, but might their Gen Z successors have figured out a better relationship with their smartphones?
Temporary workers were largely excluded from policy changes announced earlier this month after the Google walkout designed to end the use of arbitration in sexual harassment cases. The walkout organizers are trying to change that, Ellen Huet reports:
Mic.com relied heavily on Facebook to grow what it thought was an audience — and turned out to be merely traffic. After a video deal with Facebook fell through, the company laid off most of its staffers Thursday.
Activist groups Color of Change and the Leadership Conference on Civil and Human Rights wrote an angry joint letter to Facebook’s Mark Zuckerberg and Sheryl Sandberg ahead of a scheduled meeting today to discuss civil rights.
These things probably weren’t related. But they’re not … unrelated. Elsewhere in corporate philanthropy, Facebook said that it had its largest “Giving Tuesday” ever.
As Aaron Rupar says: “This isn’t normal, even for Trump.”
A Facebook bug resurfaced a bunch of old chats for some users no matter how many times they tried to close them. (I’m told by the writer Steven Levy that our 2016 chat kept popping up for him. Sorry, Steven!) Taylor Lorenz considers the strangeness of a single permanent chat thread:
These threads are just as often unnerving. Chat provides an immediate portal into your past in a way that a photo doesn’t. When you look at an old picture, you’re never remembering things the way they really were—you’re projecting your own memory of that event or day. Revisiting the same period through an old chat history is different. Chat records offer concrete evidence of the way things really felt in that moment: the embarrassing slang you used, the plans you made, the idle thoughts you shared with friends. A chat history forces you to confront a version of who you are that you probably forgot about. Part of what made Facebook users affected by the bug so uncomfortable was seeing an old version of themselves pop up without warning.
I used to wonder why Snap lost so much money every quarter. Then I read this barely coherent interview with its outgoing chief business officer. Here is his answer to how he tried to balance short- and long-term thinking at Snap:
KHAN: Yeah, look, I think what-, I think it’s really- have to appreciate what I did during the work was, that’s probably not the right thing to talk about, because of a variety of reasons, but I think if you take a step back, and look at the people who build big businesses, you know, they always really have to believe in long-term, and-, and, you know, if you don’t execute-, primarily in the technology space, the world has changed so fast, if you don’t take a long-term view, and just really manage business, from day to day, and quarter to quarter, you will miss a lot of different things. So, I think, you know, my interactions with a lot of great entrepreneurs, you know, they always really focus on long-term, so I think that’s a good thing to do, but also, at the same time, you obviously have to recognize that long-term consists of a series of short-terms, but focus-, losing focus on long-term, and who your customers are, is never the right thing to do, so I think all great entrepreneurs should focus on long-term.
Andy Baio, for one, will miss YouTube annotations when they go away next year:
YouTube annotations were supported for nearly a decade, and a significant number of creators used them to great effect — games, interactive art, education, footnotes, and corrections — and removing them fundamentally and irrevocably breaks a core part of YouTube history.
Here’s just a small piece of what we’ll lose in two months when annotations are gone. Watch them while you can.
YouTube announced its Instagram-like Stories feature late last year, Julia Alexander reports, but the feature is now rolling out to more creators.
Facebook sent a team to the University of British Columbia in Canada with free t-shirts, dogs to pet, and the opportunity to win $1,000 in exchange for downloading its murder clone of Houseparty, Bonfire. The group video chat app is now available in Canada.
Here’s a nice new accessibility feature. Ashley Carman reports:
Instagram announced today that it’s rolling out new features that will make the app easier to use for people with visual impairments. The changes will allow screen readers to describe photos, either automatically using AI or by reading custom descriptions added by users.
These descriptions, known as “alt text,” are widely used online and are now being built into Instagram. Users will be able to enter their own photo descriptions so that people using screen readers — software that describes the elements displayed on a screen — can hear it read aloud as they browse their feed, Explore page, or a profile.
Will Oremus wants to know: where is the Facebook walkout?
There is one tech behemoth whose workers have stayed oddly quiet, even as it has been pilloried by outsiders for everything from its privacy practices to its effect on democratic elections to its role in ethnic violence around the world. Employees of Facebook, possibly Silicon Valley’s most controversial company, held their peace even after a New York Times investigation that showed a pattern of company leaders conniving, dissembling, and kowtowing to conservatives as the platform’s problems mounted.
It’s time for Facebook’s workers to speak out. Past time, really. But now would be better than never.
Renee DiResta calls our current information war “a state of continuous partial conflict.” And we’re fighting the last war rather than the present one. I love her concept of “digital security theater”:
The entities best suited to mitigate the threat of any given emerging tactic will always be the platforms themselves, because they can move fast when so inclined or incentivized. The problem is that many of the mitigation strategies advanced by the platforms are the information integrity version of greenwashing; they’re a kind of digital security theater, the TSA of information warfare. Creating better reporting tools, for example, is not actually a meaningful solution for mitigating literal incitements to genocide. Malignant actors currently have safe harbor in closed communities; they can act with impunity so long as they don’t provoke the crowd into reporting them — they simply have to be smart enough to stay ahead of crowd-driven redressal mechanisms. Meanwhile, technology companies have plausible denial of complicity because they added a new field to the “report abuse” button.
Algorithmic distribution systems will always be co-opted by the best resourced or most technologically capable combatants. Soon, better AI will rewrite the playbook yet again — perhaps the digital equivalent of Blitzkrieg in its potential for capturing new territory. AI-generated audio and video deepfakes will erode trust in what we see with our own eyes, leaving us vulnerable both to faked content and to the discrediting of the actual truth by insinuation. Authenticity debates will commandeer media cycles, pushing us into an infinite loop of perpetually investigating basic facts. Chronic skepticism and the cognitive DDoS will increase polarization, leading to a consolidation of trust in distinct sets of right and left-wing authority figures – thought oligarchs speaking to entirely separate groups.
Semi-related to the above: William Davies has a long, thoughtful piece about declining trust in institutions and journalists, considering it from both technological and political perspectives. Cries of “fake news” are but a symptom of a much larger problem, he writes:
What nobody foresaw was that, when trust sinks beneath a certain point, many people may come to view the entire spectacle of politics and public life as a sham. This happens not because trust in general declines, but because key public figures – notably politicians and journalists – are perceived as untrustworthy. It is those figures specifically tasked with representing society, either as elected representatives or as professional reporters, who have lost credibility.
To understand the crisis liberal democracy faces today – whether we identify this primarily in terms of “populism” or “post-truth” – it’s not enough to simply bemoan the rising cynicism of the public. We need also to consider some of the reasons why trust has been withdrawn. The infrastructure of fact has been undermined in part by a combination of technology and market forces – but we must seriously reckon with the underlying truth of the populists’ charge against the establishment today. Too often, the rise of insurgent political parties and demagogues is viewed as the source of liberalism’s problems, rather than as a symptom. But by focusing on trust, and the failure of liberal institutions to sustain it, we get a clearer sense of why this is happening now.
And finally ...
San Francisco’s public hospital was renamed Priscilla Chan and Mark Zuckerberg San Francisco General Hospital in 2015 after the they donated $75 million to its foundation. Now Supervisor Aaron Peskin, last seen mounting a harebrained scheme to prevent companies from offering lunch to their employees, is back in the news with another empty gesture. (He’s making noise about the Definers scandal.) But:
The city has moved to rename some public structures before, such as Justin Herman Plaza on the Embarcadero, but removing Zuckerberg’s name from the hospital may trigger a dispute because of a naming agreement Zuckerberg and Chan reached with the hospital, which is owned and operated by the San Francisco Department of Public Health. The agreement, adopted by the Board of Supervisors in 2015, says the hospital is to be named the Priscilla Chan and Mark Zuckerberg San Francisco General Hospital and Trauma Center for 50 years. The $75 million gift is believed to be the single largest contribution by private individuals in support of a public hospital in the United States.
Should Peskin continue down this road? Sure, just as soon as the city resolves its ongoing housing crisis, drug crisis, and about 14,000 other issues that city residents would like to see addressed before this one.
Talk to me
Send me tips, comments, questions, and strike organizing materials: firstname.lastname@example.org.