Internal research at Meta and TikTok, sources claim, revealed that outrage-inducing content drove user engagement, leading the companies to permit more harmful material on user feeds. This occurred as the social media giants engaged in a fierce competition for user attention.
Over a dozen individuals with insider knowledge have detailed how Meta, parent company of Facebook and Instagram, and TikTok allegedly took significant risks with user safety. These risks reportedly involved issues ranging from violence and sexual blackmail to terrorism. The primary motivation, they suggest, was to gain an edge in the battle for users’ time and attention.
One engineer formerly employed by Meta described instructions from senior management to permit a greater volume of “borderline” harmful content in users’ feeds. This category of content includes misogyny and conspiracy theories. The directive was issued as Meta sought to better compete with TikTok’s rapidly growing influence.
“They sort of told us that it’s because the stock price is down,” the former Meta engineer stated, linking the decision directly to financial performance pressures.
A TikTok employee provided the BBC with rare access to internal dashboards detailing user complaints. This evidence reportedly showed that staff were directed to prioritize cases involving politicians over a series of reports concerning harmful posts that featured children. This indicates a clear shift in focus away from child protection.
The decisions to prioritize political figures, according to the TikTok staffer, were made to “maintain a strong relationship” with them. The goal was to preempt potential regulatory threats or outright bans, rather than to mitigate risks to users.
The Algorithm Arms Race Intensified by TikTok’s Rise
The whistleblowers who spoke to the BBC documentary, “Inside the Rage Machine,” provided an intimate perspective on the industry’s response to TikTok’s explosive growth. TikTok’s highly effective algorithm for recommending short videos had fundamentally altered the social media landscape, compelling rivals to scramble for parity.
Matt Motyl, a former senior researcher at Meta, commented on Instagram Reels, Meta’s competitor to TikTok. He stated that Reels was launched in 2020 without adequate safety measures in place. Internal research shared with the BBC indicated that comments on Reels contained significantly higher rates of bullying, harassment, hate speech, and violence or incitement compared to other parts of Instagram.
While Meta reportedly invested in 700 staff members to bolster Reels’ growth, safety teams were denied requests for additional specialized personnel. Specifically, they sought two staff members dedicated to child protection and ten more to support election integrity efforts, according to another former senior Meta employee.
Motyl shared dozens of documents with the BBC, which he characterized as “high-level research documents showing all sorts of harms to users on these platforms.” Among these were indications that Facebook was aware of issues stemming from its algorithms.
One internal study highlighted that the algorithm offered content creators a “path that maximizes profits at the expense of their audience’s wellbeing.” It also noted that “the current set of financial incentives our algorithms create does not appear to be aligned with our mission” to connect the world.
The study further suggested that Facebook could “choose to be idle and keep feeding users fast-food, but that only works for so long.”
In response to these allegations, Meta asserted: “Any suggestion that we deliberately amplify harmful content for financial gain is wrong.” TikTok claimed these were “fabricated claims” and stated the company had invested in technology to prevent harmful content from being viewed.
The “Black Box” of Recommendation Engines
Ruofan Ding, who worked as a machine-learning engineer on TikTok’s recommendation engine from 2020 to 2024, described the algorithms as a “black box.” He noted that their internal workings were difficult to scrutinize.
“It was hard to build systems like this that were completely safe,” Ding remarked. “We have no control of the deep-learning algorithm in itself.”
Engineers focused on the content as data points, he explained. “To us, all the content is just an ID, a different number.”
They relied on content safety teams to ensure harmful posts were removed, thereby preventing their promotion by the algorithm. Ding likened this interdependency to different teams working on separate components of a car, where one team expects another to handle the braking system effectively.
However, Ding observed that as TikTok continuously refined its algorithm to gain market share, he began to notice an increase in “borderline” content. This included problematic posts that emerged after users had spent some time browsing videos.
“Borderline content” is a term used within social media companies to describe posts that are harmful but legal. This can encompass misogynistic, racist, and sexualized content, as well as conspiracy theories.
User Experience and Algorithm Influence
Teenagers informed the BBC that the systems designed for users to flag problematic content were not functioning effectively. They continued to be recommended violent and hateful material on major social media platforms.
In an extreme instance, a teenager named Calum, now 19, stated he had been “radicalised by algorithm” from the age of 14. He reported that the algorithm showed him content that evoked outrage, leading him to adopt racist and misogynistic views.
“The videos ‘energised me, but not really in a good way,'” he said. “‘They just made me very kind of angry. It very much reflected the way I felt internally, that I was angry at the people around me.'”
Counter-terror police specialists in the UK, who routinely analyze thousands of social media posts, have observed a “normalisation” of antisemitic, racist, violent, and far-right content in recent months.
“People are more desensitised to real-world violence and they are not afraid to share their views,” one officer commented.
Inside TikTok’s Trust and Safety Team
Over several months in 2025, the BBC communicated with a member of TikTok’s trust and safety team, referred to as “Nick” for anonymity. They were shown the company’s internal dashboard, which detailed the cases his team was handling and their response strategies.
“If you’re feeling guilty on a daily basis because of what you’re instructed to do, at some point you can decide, should I say something?” Nick questioned, reflecting on the ethical toll.
He stated that the sheer volume of cases being reviewed made it exceedingly difficult to ensure user safety, particularly for teenagers and children who were left at heightened risk. Nick suggested that recent cuts and reorganizations of moderation teams, including the replacement of some human roles with AI technology, had diminished the company’s capacity to effectively address such content.
Material linked to “terrorism, sexual violence, physical violence, abuse, trafficking” was reportedly increasing, according to the whistleblower.
The reality of the app’s recommendations and the actions taken against harmful content, he added, is “very different in a lot of aspects to what the sites are saying” publicly.
Nick provided evidence to the BBC demonstrating how TikTok rated certain relatively minor cases involving politicians as higher priority for the safety team’s review than several cases involving harm to teenagers.
In one instance, a political figure who had been mocked by being compared to a chicken was prioritized over a 17-year-old reporting cyberbullying and impersonation in France. Another case involved a 16-year-old in Iraq who complained about sexualized images purportedly of her being shared on the app; this was assigned a lower priority.
Regarding the Iraq case, Nick observed, “If you look at the country where this report comes from, it’s very high risk because it’s a minor and it involves sexual blackmail and then you can see the priority here. The urgency is not high.”
Nick also showed examples where certain posts encouraging people to join terror groups or commit crimes had not been designated as top priority.
When the trust and safety team requested to prioritize cases involving young people over political ones, the whistleblower stated they were instructed not to, and to continue handling cases according to their assigned ranking.
Nick posited that the reason for this prioritization was that the company, in his view, cared less about children’s safety than maintaining a “strong relationship” with politicians and governments. This approach aimed to avoid regulatory intervention or bans that could negatively impact its business.
Nick mentioned that when he and other staff members raised these concerns with management, their feedback was not met with receptiveness. He attributed this to management not being “exposed to this content on a day-to-day basis.”
The trust and safety employee offered blunt advice to parents whose children use TikTok: “Delete it, keep them as far away as possible from the app for as long as possible.”
TikTok rejected the notion that political content was prioritized over the safety of young people, calling the claim that it “fundamentally misrepresents the way their moderation systems operate.”
The company stated that Nick’s team is part of a broader safety system with multiple teams responsible for reviewing content complaints. TikTok elaborated, “Specialist workflows for certain issues do not result in the deprioritization of child safety cases, which are handled by dedicated teams within parallel review structures.”
A TikTok spokesperson commented that the criticisms “ignore the reality of how TikTok enables millions to discover new interests, find community, and supports a thriving creator economy.”
The company also highlighted that accounts for teens have over 50 preset safety features and settings that are automatically enabled. Additionally, it invests in technology to prevent harmful content from being viewed, maintains strict recommendation policies, and provides features for users to customize their experiences.
The Race to Catch Up
The algorithm arms race intensified in 2020 with the introduction of Instagram Reels, a direct response to TikTok’s global surge in popularity, particularly during the COVID-19 pandemic.
Matt Motyl, who worked as a senior researcher at Facebook and its successor, Meta, from 2021, described Reels as the company’s effort to “mimic” the “unique product” that TikTok had launched.
His role between 2019 and 2023 involved conducting “large-scale experiments on sometimes as many as hundreds of millions of people”—many of whom were unaware this was occurring—to test content ranking in user feeds.
“Meta’s products are used by north of three billion people and the more time they can keep you on there, the more ads they sell, the more money they make. But it’s very important that they get this stuff right, because when they don’t, really bad things happen,” he stated.
Regarding Reels specifically, Motyl noted that the strategy was to proceed as rapidly as possible, irrespective of the impact on users. He acknowledged a “common trade-off between protecting people from harmful content and engagement.”
According to a research paper Motyl shared with the BBC, Meta struggled to prevent harm on Reels following its launch. The paper suggested that Reels posts featured a higher prevalence of harmful comments than posts on the main Instagram feed: 75% more bullying and harassment, 19% more hate speech, and 7% more violence and incitement.
He described a “power imbalance” where safety staff required agreement from the teams managing Reels to implement new products or features aimed at improving user safety. The Reels teams, Motyl explained, had “incentives to not let those products launch because toxic stuff gets more engagement than non-toxic.”
Brandon Silverman, whose social media monitoring tool Crowdtangle was acquired by Facebook in 2016, participated in senior-level discussions during that period. He characterized CEO Mark Zuckerberg as “very paranoid” about competition.
“When he feels like there are potential competitive forces there’s no amount of money that is too much,” Silverman said.
He observed that during this time, safety teams faced challenges in gaining approval for recruiting small numbers of staff, while the company’s primary focus was on expanding Reels. “There was another team that went, oh, we just got 700 for Instagram Reels. I was like, OK yeah,” he recalled.
A former engineer at Meta, identified as “Tim” for anonymity, stated that as the company aimed to compete with TikTok, more borderline harmful content was allowed on the platform. His team had been dedicated to reducing this content until the “business positioning” shifted.
“You’re losing to TikTok and therefore your stock price must suffer. People started becoming paranoid and reactive and they were like, let’s just do whatever we can to catch up. Where can we get like 2%, 3% revenue for the next quarter?” Tim explained.
He indicated that the decision to cease limiting content that was potentially harmful but not illegal—and which users were engaging with—was made by a senior vice-president at Meta, who Tim believed reported directly to Mark Zuckerberg.
Amplifying Outrage
Around the time Facebook maintained it was merely a “mirror to society,” internal documents provided to the BBC by senior researcher Matt Motyl revealed the company’s awareness that it was amplifying content that provoked anger and even incited harm.
These documents detailed how sensitive content—which could include material touching on people’s moral beliefs or posts inciting violence—was more likely to elicit a strong reaction and engagement on the platform, particularly if it generated outrage.
“Given the disproportionate engagement, our algorithms presume that users like that content and want more of it,” the study stated.
Silverman commented that Meta’s leadership seemed initially uncertain on how to address toxic content on the platform, and that there was a period where the company appeared “genuinely introspective.”
However, he noted that their stance “began to calcify into a sort of defensiveness.” Their attitude, he said, was that “we’re not responsible for all of polarisation in society.”
“Nobody’s saying you’re responsible for all polarisation. We’re just saying you contribute to it, and probably in ways where you don’t have to. If you just made a few changes, you might not contribute to it as much,” Silverman countered.
A Meta spokesperson denied the whistleblowers’ assertions. “The truth is, we have strict policies to protect users on our platforms and have made significant investments in safety and security over the last decade,” the spokesperson stated.
The company also maintained it had “made real changes to protect teens online,” including the introduction of a new Teen Accounts feature “with built-in protections and tools for parents to manage their teens’ experiences.”