Nearly half of girls and a third of all teenagers saw suicide, self-harm and eating disorder content on social media in a week, a study shows.

The Molly Rose Foundation (MRF) research found that 47% of girls aged 13 to 17 encountered high-risk content during a seven-day period.

Only slightly fewer teens are seeing harmful content now (34%) than just before new safety measures came into force last summer (37%), the study found.

The charity, set up in memory of 14-year-old Molly Russell, who took her own life in 2017 after viewing harmful content online, said the study showed children are still facing “a tsunami of harmful content”.

The findings are based on a survey of 1,825 UK children aged 13 to 17 conducted by MEL Research in April 2026 with support from the PSHE Association.

They also showed that children with low wellbeing (57%) and those with special educational needs (40%) were found to be at even greater risk of seeing the content.

The new protection that came into force in July last year included age checks to prevent children from accessing pornography and other harmful content.

The changes also require platforms to ensure algorithms do not push content about subjects such as self-harm and eating disorders towards children.

Actions that could be taken against firms that fail to adhere to the new codes include fines of up to £18m or 10% of qualifying worldwide revenue, whichever is greater, and court orders blocking access in the UK.

Keir Starmer is expected to announce next week a ban preventing under-16s from accessing harmful social media sites, following the government’s consultation on what restrictions should be introduced.

The consultation received about 116,000 responses, making it the second-largest government consultation in history.

Ian Russell, Molly’s father, said: “It is shocking but sadly unsurprising that millions of teens continue to be shown appalling suicide, self-harm and depression content by out-of-control algorithms.

“We’ve repeatedly warned that weak implementation of the Online Safety Act would leave preventable harm unchecked, and regrettably, this research endorses these warnings.

“Keir Starmer now needs to make a choice 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.”

A Downing Street spokesperson said: “We have undertaken a thorough consultation and will set out next steps in due course.

“The prime minister has been clear that the status quo is not good enough and we need to do more to protect children.

“This is not about politics – it is about protecting children.”

Meanwhile, the Scottish government has urged Starmer to do “more to protect children and young people from online harm”.

The Scottish minister for children, Siobhian Brown, is due to meet the UK AI and online safety minister, Kanishka Narayan, on Sunday.

Brown said she would “call for urgent action to protect young people” during the meeting.

“We’ve been consistently pushing the UK government to act, given the powers to control online safety are reserved to Westminster,” she said.

“We know there are steps that the UK government could be taking now – from banning the use of social media for children and using Ofcom’s powers to force social media firms to act.”

An Institute for Public Policy Research survey of more than 2,000 adults found 51% trust parents to decide which platforms are appropriate, 49% trust an independent regulator, 22% trust schools, 16% trust technology companies and 15% trust government ministers.

The polling, conducted by YouGov, also found 44% support banning under-16s from social media, while 39% prefer tighter regulation.

Brown said she wanted to see the introduction of a social media levy, charging social media platforms a fee that could be reinvested into programmes for young people’s mental health.

“In Scotland, our aim is to tackle this issue as a public health matter, recognising that there is a spectrum of harm that can be caused from the absorption of hateful and harmful content and unhealthy use of online services.”