Millions of teens shown appalling content by out-of-control algorithms
For years, the death of fourteen-year-old Molly Russell has stood as a warning about what algorithms can do to a vulnerable young mind. Now, new research from the foundation bearing her name reveals that nearly half of British girls aged thirteen to seventeen encountered content about suicide, self-harm, and eating disorders on social media within a single week — and that landmark safety legislation introduced nine months ago has done almost nothing to change that. The United Kingdom stands at a crossroads between a sweeping ban on social media for under-sixteens and the harder, slower work of forcing the platforms themselves to stop profiting from harm.
- Nine months after the Online Safety Act's new protections took effect, harmful content exposure among teenagers has dropped by barely three percentage points — a near-negligible shift that has shaken confidence in the law's enforcement.
- The algorithms are not merely failing to protect the most vulnerable teenagers; research suggests they are actively directing self-harm and suicide content toward children with low wellbeing and special educational needs at even higher rates.
- Prime Minister Starmer is expected to announce a full social media ban for under-sixteens, backed by a record-breaking government consultation that drew 116,000 responses — the second largest in British history.
- Molly Russell's father, Ian Russell, warns that a blanket ban risks becoming a politically convenient gesture that sidesteps the product safety failures that killed his daughter, calling on the government to choose accountability over expediency.
- Scotland is pushing beyond Westminster's approach, seeking both stricter restrictions and a social media levy that would redirect platform revenues into youth mental health services, framing the crisis as a public health emergency rather than a regulatory puzzle.
When the Molly Rose Foundation surveyed 1,825 British teenagers in April, it found that 47 percent of girls aged thirteen to seventeen had encountered content about suicide, self-harm, and eating disorders on social media within a single week. The figure landed as a particular blow because it came nine months after sweeping new safety protections had been introduced — protections that, by the numbers, had barely moved the needle.
The foundation carries the name of Molly Russell, who was fourteen when she died by suicide in 2017 after viewing harmful content online. Her father, Ian Russell, has spent years pushing for change. When the new data arrived, he called it shocking but unsurprising — the words of someone who has watched his warnings go partially unheeded. Millions of teenagers, he said, continue to be shown appalling content by out-of-control algorithms.
The July 2025 protections had real teeth on paper: mandatory age verification, algorithmic restrictions on self-harm content, and fines reaching £18 million or 10 percent of global revenue. Yet the harm continued. Worse, the research showed that vulnerable groups faced disproportionate exposure — 57 percent of children with low wellbeing encountered such content, as did 40 percent of those with special educational needs. The algorithms were not simply failing to protect the most fragile young people; they appeared to be targeting them.
Prime Minister Keir Starmer is expected to announce a ban on social media access for anyone under sixteen — a proposal that drew 116,000 consultation responses, making it the second-largest government consultation in British history. But Ian Russell has posed the harder question: is a blanket ban the right answer, or does it let the platforms off the hook? He has urged Starmer to choose between political expediency and finally confronting the product safety failures that cost his daughter her life.
Scotland is pushing further still, with minister Siobhian Brown calling not only for tighter restrictions but for a social media levy — fees charged to platforms and reinvested in youth mental health. Public polling reflects a nation genuinely divided: 44 percent support an outright ban, 39 percent prefer stronger regulation, and only 16 percent trust technology companies to make the right calls. The country is being asked to decide whether to restrict access entirely, or to finally force the platforms themselves to change.
Nearly half of all girls between thirteen and seventeen saw content about suicide, self-harm, and eating disorders on social media within a single week. This is what the Molly Rose Foundation found when it surveyed 1,825 British teenagers in April of this year. The figure—47 percent—arrived as a particular blow because it suggested that the safety measures introduced just nine months earlier had barely moved the needle. Before those protections took effect last summer, 37 percent of all teenagers reported encountering such material. Now it sits at 34 percent. The gap is negligible.
The Molly Rose Foundation exists because of a girl named Molly Russell, who was fourteen when she died by suicide in 2017 after spending time online viewing harmful content. Her father, Ian Russell, has spent years pushing for change. When he saw these new numbers, he called them shocking but unsurprising—a phrase that carries the weight of someone who has watched his warnings go partially unheeded. "Millions of teens continue to be shown appalling suicide, self-harm and depression content by out-of-control algorithms," he said, and the research bore him out.
The new protections that came into force in July 2025 were meant to be substantial. They required age verification systems to keep children away from pornography and other dangerous material. They mandated that algorithms stop actively pushing content about self-harm and eating disorders toward young users. Platforms that failed to comply faced fines up to £18 million or 10 percent of their global revenue—whichever was larger—plus potential court orders blocking their access to British users. On paper, the teeth were there. In practice, the harm continued.
The research revealed something darker still: certain groups of teenagers face disproportionate exposure. Children reporting low wellbeing encountered this content at a rate of 57 percent. Those with special educational needs faced it at 40 percent. The algorithms, it seemed, were not just failing to protect vulnerable young people—they were actively targeting them.
Prime Minister Keir Starmer is expected to announce next week a ban preventing anyone under sixteen from accessing social media platforms altogether. The government's consultation on this proposal drew 116,000 responses, making it the second-largest government consultation in British history. The public appetite for action is clear. But Ian Russell has posed a harder question: Is a blanket ban the right move, or does it sidestep the real problem? "Keir Starmer now needs to make a choice," he said, "between a politically expedient blanket ban that the evidence says will quickly fail or finally addressing the product safety risks that cost my daughter Molly's life."
Downing Street responded with measured language. "The status quo is not good enough," a spokesperson said. "We need to do more to protect children." Scotland's government, meanwhile, is pushing further. Minister Siobhian Brown is scheduled to meet with the UK's AI and online safety minister this Sunday, and she plans to call for urgent action. She wants to see not just restrictions but a social media levy—a fee charged to platforms that could be reinvested into mental health programs for young people. She frames the issue as a public health crisis, not merely a regulatory one.
Public opinion, according to polling by the Institute for Public Policy Research, is divided on who should decide what children see online. Fifty-one percent trust parents to make those calls. Forty-nine percent trust an independent regulator. Only 16 percent trust technology companies, and just 15 percent trust government ministers. When asked about solutions, 44 percent support an outright ban on under-16s accessing social media, while 39 percent prefer tighter regulation. The consultation responses and these numbers suggest the country is at a crossroads: either restrict access entirely, or finally force the platforms themselves to change how they operate.
Citas Notables
Millions of teens continue to be shown appalling suicide, self-harm and depression content by out-of-control algorithms.— Ian Russell, father of Molly Russell
The status quo is not good enough and we need to do more to protect children.— Downing Street spokesperson
La Conversación del Hearth Otra perspectiva de la historia
Why did the new safety measures have so little effect? The fines are enormous—up to £18 million.
Because enforcement is slow and platforms are vast. A fine issued months after the fact doesn't stop an algorithm running today. The rules say platforms must prevent harmful content from being pushed to children, but proving they violated that rule takes investigation, evidence, legal process. By then, millions of teenagers have already seen the material.
The study found that girls are hit much harder than boys—47 percent versus roughly a third overall. Why?
The algorithms learn what keeps people engaged. Self-harm and eating disorder content tends to create loops of engagement, especially among girls. Once the system detects that pattern, it amplifies it. It's not malice; it's optimization without conscience.
Ian Russell mentions a "blanket ban" as potentially ineffective. What does he mean?
He's saying that if you simply ban under-16s from social media, they'll find workarounds—fake ages, VPNs, other platforms. The real problem isn't access; it's what the platforms do with the attention once you're there. A ban treats the symptom, not the disease.
Scotland wants a social media levy. How would that actually help?
Money flowing into mental health services for young people. It's an acknowledgment that the platforms are causing harm and should fund the treatment. It's also a way to pressure platforms without a blanket restriction—you pay, or you face consequences.
The consultation got 116,000 responses. What does that number tell you?
People are frightened. Parents are frightened. The second-largest government consultation in history means this isn't a niche concern—it's a crisis that's reached the mainstream. The government can't ignore it, and they know it.