The Washington Post Democracy Dies in Darkness

The Technology 202: Facebook's new emphasis on groups could leave it more vulnerable to disinformation

May 1, 2019 at 9:02 a.m. EDT

Ctrl + N

Facebook is making communication with smaller communities “the heart” of its biggest redesign in years. But that change could leave the service even more vulnerable to disinformation artists.

Facebook chief executive Mark Zuckerberg announced yesterday that communications on its Groups feature will be as central as connecting with Facebook “friends.” They’ll be front and center in the company's app as Facebook undergoes a makeover that Zuckerberg says signals the company’s plans to reorient toward privacy. 

But skeptics warn that Facebook groups are a preferred tool for trolls, spammers and other bad actors seeking to stoke divisions or blast fake news. They’ve been leveraged by Russian trolls ahead of U.S. elections and exploited by hoaxers spreading false information about vaccinations.

“This may be good for Facebook’s business, but I’m not sure it’s good for society at large,” said Paul Barrett, deputy director of New York University’s Stern Center for Business and Human Rights.

This strategy could put Facebook in a serious bind as policymakers in Washington and around the world want the company to beef up its defenses against disinformation, especially ahead of key elections. Putting Groups front and center could lead to political headaches down the line as lawmakers scrutinize the role disinformation on social media can play in influencing election outcomes or public health decisions.

The shift to communities means Facebook may put more of the onus on users to play a bigger role in policing content on the website. Already, group administrators set the rules and monitor the communities they administrate on Facebook. If such groups become a more prominent feature on the site, these users could be a bigger line of defense against disinformation on the service. 

I think it does shift a great deal more responsibility onto the shoulders of the group administrators,” Barrett told me. 

But not all group administrators have good intentions — which could exacerbate safety problems. “The thing that makes me nervous is the formation of groups for malign purposes — situations where the admins might be part of the problem, not the cops on the beat,” Barrett said. 

Facebook has a business incentive to make communities a more prominent part of the experience. As Zuckerberg noted in his keynote at the F8 developers conference yesterday, communications with smaller groups of friends — through messaging, ephemeral stories and Groups — are the fastest-growing ways people are communicating online. 

Zuckerberg described the positive benefits of groups in building digital communities. He touted groups that connect military families moving to new cities, or one that connects runners who care about the environment and want to coordinate jogs where they pick up trash. 

But he didn't mention some of the more sinister ways that Facebook Groups — which can have varying levels of secrecy depending on the administrator settings — can be leveraged.

For instance, special counsel Robert S. Mueller III's indictment of a Russian troll farm in early 2018 exposed how the actors seeking to influence the election used groups on the platform to target Americans on politically divisive issues. BuzzFeed reported how Facebook groups were used to troll people who thought they were seeking information about the 2018 March for Our Lives rally in favor of gun control. My colleague Lena Sun detailed how a closed Facebook group was ground zero for a plot to troll a Pittsburgh doctor's office urging parents to vaccinate their children. 

Zuckerberg did acknowledge in his keynote that the company is prioritizing safety as it makes this shift. 

“We are very focused on safety here across groups,” Zuckerberg said. “While we're recommending groups for people to join, we're very focused on making sure that our recommendation and discovery services aren't highlighting groups where people are repeatedly sharing misinformation or harmful content. We're working hard to remove groups if they exist primarily to violate our policies or do things that are dangerous.”

But Barrett frets that the groups will simply replicate the same disinformation problems on the platform currently. "My reaction was one of great skepticism and concern that in an effort to hold onto user volume, keep as many users as possible, that Facebook could end up just going down a different path that leads to same problems we’ve seen before," Barrett said.

Facebook's decision to elevate groups follows the company's announcement that it would shift to more private and encrypted forms of communication. As we previously covered in The Technology 202, the broader strategy of moving toward encryption comes with clear safety trade-offs. 

Zuckerberg told my colleague Elizabeth Dwoskin in an interview this week that the company is working on ways to identify coordinated activity from bad actors even when it can't see the messages, “but there's no way that I can sit here and tell you that we can do 100 percent as well if we eliminate one of the big tools that we have in people's messages.”

BITS, NIBBLES AND BYTES

BITS: Facebook's redesign represents the company's efforts to present a new face to the world after a year of controversies, my colleague Elizabeth reports from Facebook's developer conference in Menlo ParkBut beyond these cosmetic changes, Zuckerberg is still debating how to respond to regulation and the potential tradeoffs that could come with the company's plans to reorient toward privacy and encryption. 

“The next five years at least, maybe even the next 10 years, is building out the private platforms with the richness that the public platforms have had to date,” Zuckerberg told Elizabeth in an interview the night before the conference. “That needs to get done with some amount of different infrastructure and different policies, and to some degree, different values, than in building out these public spaces.”

As the company moves more toward private messaging, it hasn't yet figured out how to make money off that behavior. Ninety-nine percent of the company’s revenue comes from targeted advertising, which is based on information the company collects in part from people's public posts. When the company encrypts people's messages end-to-end, even Facebook won't be able to read them -- which could disrupt its advertising practices. 

Zuckerberg doesn't know how much profit the company will derive from encrypted messaging, but he's not worried. 

“I don’t know how good of a business it will be, but I am confident it will be good, and we will be fine,” he said in the interview.

A sheriff's office in Oregon is using Amazon Rekognition to help track down criminal suspects. Is this a dangerous precedent against privacy? (Video: Drew Harwell, Jhaan Elker/The Washington Post)

NIBBLES: The Washington County Sheriff's Office in Oregon became the first law enforcement agency in the country to use Amazon's facial recognition tool, Rekognition. It also became ground zero for a battle over the controversial practice of using nascent technology in policing, my colleague Drew Harwell reports. (Amazon founder Jeff Bezos also owns The Post.)

The facial recognition technology has supercharged the abilities of police in the region, allowing them to swiftly see if an image matches up to one of the 30,000 mug shots taken at its county jail since 2001. Even grainy pictures captured on security cameras or a deputy's smartphone can quickly provide a link to a person's name, address and family. More than 1,000 searches were logged last year, and sometimes the results were used to find a person's Facebook page, visit their home or make an arrest. 

"Defense attorneys, artificial-intelligence researchers and civil rights experts argue that the technology could lead to the wrongful arrest of innocent people who bear only a resemblance to a video image," Drew wrote. "Rekognition’s accuracy is also hotly disputed, and some experts worry that a case of mistaken identity by armed deputies could have dangerous implications, threatening privacy and people’s lives."

It's not known how accurate the technology has been during the first 18 months of testing. 

"Deputies don’t have to note in arrest reports when a facial-recognition search was used, and the exact number of times it has resulted in an arrest is unclear," Drew wrote. "Sheriff’s officials said the software has led to dozens of arrests for theft, violence or other crimes, but a public-records request turned up only nine case reports in which facial recognition was mentioned."

Amazon’s guidelines for law enforcement say officials should use the technology's findings only when the system is 99 percent confident in a match. But deputies in the region don't see that percentage when they use the tool. Amazon updated the language of its guidelines in response to Drew's story to say law enforcement officials should manually review all matches before detaining a suspect and that the search “shouldn’t be used as the sole determinant for taking action.”

BYTES: A cancer center in Tampa warned lawmakers that robocalls are a threat to both doctors and patients as they urged Congress to take action, my colleague Tony Romm wrote

The spike in robocalls has proven especially troubling for hospitals, treatment centers and other health-related organizations, said Dave Summitt, the chief information security officer at the H. Lee Moffitt Cancer Center & Research Institute. Testifying at a before the House Energy and Commerce Committee as lawmakers consider a bill to tackle robocalls, he said the institute received roughly 6,600 external calls over a recent 90-day period that “were either malicious intent, or identified themselves as someone they’re not,” using numbers that appeared to be coming from within the cancer center itself.

“When you’re sitting here, and you’re in a health-care situation, and you’re seeing a phone call coming from inside the organization, you’re going to pick the thing up,” Summit said. “If they happen to get a hold of one of our patients … they are absolutely going to answer that phone.”

“They are making money, and they’re doing it on the backs of patients and other consumers,” Summitt added. “And in the process, they’re hurting us very badly.”

Lawmakers are considering updating decades-old robocall rules, and yesterday's hearing focused on seven possible changes, including a bill that would toughen penalties for scammers. Lawmakers from both parties expressed optimism they could come together on the issue. 

RANT AND RAVE

Delivering his keynote at the developers' conference, Zuckerberg stood under a screen that said “The future is private.” Journalists and tech experts responded on Twitter to the shift in the company's strategy. 

Kevin Roose of the New York Times noted the company's tune has changed since its early days:

Will Oremus, a senior writer for OneZero at Medium, noted the company's big privacy promises didn't add up with its new commitments to advertisers: 

The Wall Street Journal's Tomio Geron joked:

BuzzFeed News reporter Pranav Dixit noted that Zuckerberg's attempt to make light of the company's recent privacy mishaps didn't land:

PRIVATE CLOUD

-- Tech news from the private sector: 

Former Google CEO Eric Schmidt will leave Alphabet’s board after 18 years (The Verge)

Google workers are holding a sit-in May 1 to protest alleged retaliation (CNET)

Apple says its services business grew as fast as iPhone sales fell (Reed Albergotti)

Would you confess your secret crush to Facebook? (Lisa Bonos)

PUBLIC CLOUD

-- Tech news from the public sector: 

WikiLeaks Founder Julian Assange Sentenced to 50 Weeks in Jail (Wall Street Journal)

Social media remains key to Venezuela’s opposition, despite efforts to block it (Drew Harwell and Mariana Zuñiga)

Judge Approves Deal Spelling Out Oversight of Elon Musk’s Tweets (Wall Street Journal)

#TRENDING

-- Tech news generating buzz around the Web:

Slack vocab game that gave away money for no real reason is getting shut down (The Verge)

Workers Love AirPods Because Employers Stole Their Walls (The Atlantic)

The hiring spreadsheet and the clash at The Markup (Columbia Journalism Review)

WIRED IN

Senate Majority Leader Mitch McConnell (R-Ky.) sidestepped three questions on April 30 about whether congressional subpoenas were mandatory. (Video: The Washington Post)