Unredacted Meta documents reveal ‘historical reluctance’ to protect children

Internal Meta documents about child safety have been unsealed as part of a lawsuit filed by the New Mexico Department of Justice against both Meta and its CEO, Mark Zuckerberg. The documents reveal that Meta not only intentionally marketed its messaging platforms to children, but also knew about the massive volume of inappropriate and sexually explicit content being shared between adults and minors. 

The documents, unsealed on Wednesday as part of an amended complaint, highlight multiple instances of Meta employees internally raising concerns over the exploitation of children and teenagers on the company’s private messaging platforms. Meta recognized the risks that Messenger and Instagram DMs posed to underaged users, but failed to prioritize implementing safeguards or outright blocked child safety features because they weren’t profitable. 

In a statement to TechCrunch, New Mexico Attorney General Raúl Torrez said that Meta and Zuckerberg enabled child predators to sexually exploit children. He recently raised concerns over Meta enabling end-to-end encryption protection for Messenger, which began rolling out last month. In a separate filing, Torrez pointed out that Meta failed to address child exploitation on its platform, and that encryption without proper safeguards would further endanger minors. 

“For years, Meta employees tried to sound the alarm about how decisions made by Meta executives subjected children to dangerous solicitations and child exploitation,” Torrez continued. “Meta executives, including Mr. Zuckerberg, consistently made decisions that put growth ahead of children’s safety. While the company continues to downplay the illegal and harmful activity children are exposed to on its platforms, Meta’s internal data and presentations show the problem is severe and pervasive.” 

Originally filed in December, the lawsuit alleges that Meta platforms like Instagram and Facebook have become “a marketplace for predators in search of children upon whom to prey,” and that Meta failed to remove many instances of child sexual abuse material (CSAM) after they were reported on Instagram and Facebook. Upon creating decoy accounts purporting to be 14-year-olds or younger, the New Mexico DOJ said Meta’s algorithms turned up CSAM, as well as accounts facilitating the buying and selling of CSAM. According to a press release about the lawsuit, “certain child exploitative content is over ten times more prevalent on Facebook and Instagram than it is on Pornhub and OnlyFans.”

In response to the complaint, a Meta spokesperson told TechCrunch, “We want teens to have safe, age-appropriate experiences online, and we have over 30 tools to support them and their parents. We’ve spent a decade working on these issues and hiring people who have dedicated their careers to keeping young people safe and supported online. The complaint mischaracterizes our work using selective quotes and cherry-picked documents.”

The unsealed documents show that Meta intentionally tried to recruit children and teenagers to Messenger, limiting safety features in the process. A 2016 presentation, for example, raised concerns over the company’s waning popularity among teenagers, who were spending more time on Snapchat and YouTube than on Facebook, and outlined a plan to “win over” new teenage users. An internal email from 2017 notes that a Facebook executive opposed scanning Messenger for “harmful content,” because it would be a “competitive disadvantage vs other apps who might offer more privacy.” 

The fact that Meta knew that its services were so popular with children makes its failure to protect young users against sexual exploitation “all the more egregious,” the documents state. A 2020 presentation notes that the company’s “End Game” was to “become the primary kid messaging app in the U.S. by 2022.” It also noted Messenger’s popularity among 6 to 10-year-olds. 

Meta’s acknowledgement of the child safety issues on its platform is particularly damning. An internal presentation from 2021, for example, estimated that 100,000 children per day were sexually harassed on Meta’s messaging platforms, and received sexually explicit content like photos of adult genitalia. In 2020, Meta employees fretted over the platform’s potential removal from the App Store after an Apple executive complained that their 12-year-old was solicited on Instagram. 

“This is the kind of thing that pisses Apple off,” an internal document stated. Employees also questioned whether Meta had a timeline for stopping “adults from messaging minors on IG Direct.” 

Another internal document from 2020 revealed that the safeguards implemented on Facebook, such as preventing “unconnected” adults from messaging minors, did not exist on Instagram. Implementing the same safeguards on Instagram was “not prioritized.” Meta considered allowing adult relatives to reach out to children on Instagram Direct a “big growth bet” — which a Meta employee criticized as a “less than compelling” reason for failing to establish safety features. The employee also noted that grooming occurred twice as much on Instagram as it did on Facebook. 

Meta addressed grooming in another presentation on child safety in March 2021, which stated that its “measurement, detection and safeguards” were “more mature” on Facebook and Messenger than on Instagram. The presentation noted that Meta was “underinvested in minor sexualization on IG,” particularly in sexual comments left on minor creators’ posts, and described the problem as a “terrible experience for creators and bystanders.” 

“Child exploitation is a horrific crime and online predators are determined criminals,” a Meta spokesperson told TechCrunch. “We use sophisticated technology, hire child safety experts, report content to the National Center for Missing and Exploited Children, and share information and tools with other companies and law enforcement, including state attorneys general, to help root out predators. In one month alone, we disabled more than half a million accounts for violating our child safety policies.”

Meta has long faced scrutiny for its failures to adequately moderate CSAM. Large U.S.-based social media platforms are legally required to report instances of CSAM to the National Center for Missing & Exploited Children (NCMEC)’s CyberTipline. According to NCMEC’s most recently published data from 2022, Facebook submitted about 21 million reports of CSAM, making up about 66% of all reports sent to the CyberTipline that year. When including reports from Instagram (5 million) and WhatsApp (1 million), Meta platforms are responsible for about 85% of all reports made to NCMEC. 

This disproportionate figure could be explained by Meta’s overwhelmingly large user base, constituting over 3 billion daily active users. A Meta spokesperson said that these numbers are a result of proactive detection. Still, in response to much research, international leaders have argued that Meta isn’t doing enough to mitigate these millions of reports. In June, Meta told the Wall Street Journal that it had taken down 27 networks of pedophiles in the last two years, yet researchers were still able to uncover numerous interconnected accounts that buy, sell and distribute CSAM. In the five months after the Journal’s report, it found that Meta’s recommendation algorithms continued to serve CSAM; though Meta removed certain hashtags, other pedophilic hashtags popped up in their place.

Meanwhile, Meta is facing another lawsuit from 42 U.S. state attorneys general over the platforms’ impact on children’s mental health. 

“We see that Meta knows that its social media platforms are used by millions of kids under 13, and they unlawfully collect their personal info,” California Attorney General Rob Bonta told TechCrunch in November. “It shows that common practice where Meta says one thing in its public-facing comments to Congress and other regulators, while internally it says something else.”

Update, 1/17/24, 11:30 PM ET with comments from Meta.