Social Media and Self-Harm: New Study Reveals Urgent Need for Better Content Control and Safer Spaces

Social media and self-harm have become two of the most tangled topics in youth mental health today, and a major new study is forcing a rethink of how we talk about both. Researchers at Swansea University’s National Centre for Suicide Prevention and Self-Harm Research surveyed more than 5,000 people and interviewed 17 individuals with lived experience of self-harm who had also encountered self-harm content online. Their goal was simple but overdue: ask the people actually affected what works, what doesn’t, and what safety measures feel like from the inside.

The findings complicate the usual narrative. Blanket bans and heavy-handed content removal might sound like an easy fix, but the study found they can backfire, cutting people off from peer support, recovery stories, and spaces where they finally feel understood. At the same time, almost everyone wanted more personal control over what shows up in their feed. This isn’t a story about social media being simply good or bad. It’s about the messy middle ground where platforms, policymakers, and users need to build something smarter than a blunt filter. In this article, we’ll break down what the study found, why current restrictions aren’t hitting the mark, and what a genuinely safer online space could look like.

What the Swansea University Study Actually Found

The research, published in the Journal of Medical Internet Research (JMIR), is described by its authors as the first study to directly gather the perspectives of people with a history of self-harm on the safety features platforms have rolled out. That distinction matters. Most previous policy decisions around self-harm content were made by platform executives, lawmakers, or advocacy groups, not by the people who actually use these spaces to cope, connect, or recover.

Here’s what stood out from the data:

• Half of respondents felt that additional restrictions introduced by platforms since 2019 made social media safer overall.
• 34.9% said the changes altered what content they saw, while 9.9% said it changed what they were willing to post.
• 79.4% believed restrictions could leave someone feeling bad or isolated, particularly when they cut off access to recovery-focused communities.
• 21.1% had a post removed or censored because it showed visible self-harm scars, and 90% of those affected described the experience as harmful.
• 87.8% of respondents wanted more personal control over the content they see.
• More than 90% said they’d click through a generic content warning, but only about 40% would click a warning specifically labeled for self-harm or suicide content, since the specificity let them make an informed choice.

Dr. Amanda Marchant, the study’s first author, put it plainly: restrictions may reduce exposure to harmful posts, but they can also unintentionally strip away recovery narratives and community support. Her point isn’t that safety measures are pointless. It’s that they need constant evaluation, not a one-and-done rollout.

Why Blanket Bans Aren’t the Simple Fix They Seem

There’s been growing political pressure, especially in places like Australia, to restrict social media access for children and teenagers outright. It’s an understandable instinct. If a platform is a source of harm, remove the access. But the Swansea researchers point to a complication: early evaluations of Australia’s under-16 ban found that roughly eight in ten children were still finding ways onto social media anyway.

This is the core tension the study surfaces. Bans and heavy content restrictions look decisive on paper, but they tend to produce a few unintended consequences:

1. Reduced access to supportive spaces. Many young people rely on peer communities online, especially when they don’t feel safe or comfortable reaching out to a traditional support service.
2. Increased isolation. Cutting someone off from a community they depend on doesn’t remove the underlying distress; it just removes the outlet.
3. Displacement to darker corners of the internet. Professor Ann John, the study’s senior author and Director of the NCSR, warned that bans can push young people toward “more concealed spaces on the internet” that are far harder to monitor or moderate.
4. Breakdown in communication with trusted adults. When social media use becomes something to hide, young people become less likely to talk about their online experiences with parents, teachers, or clinicians.

None of this means restrictions are useless. It means restrictions designed without input from the people they affect tend to solve the visible problem while creating a harder-to-see one.

The Role of Content Warnings and Why Specificity Matters

One of the more actionable findings from the research involves content warnings themselves. Respondents were far more willing to engage with a generic warning label than a specific self-harm or suicide-related one. On the surface, that might look like people avoiding the very warnings meant to protect them. But the interviews suggest something different: it’s about informed choice.

When a warning clearly states what kind of content lies behind it, users can decide for themselves whether they’re in the right headspace to view it. A vague or generic label removes that agency. This lines up with what mental health researchers have long argued: self-harm and suicide prevention strategies work best when they respect a person’s autonomy rather than trying to shield them from every possible trigger. The Centers for Disease Control and Prevention notes that effective prevention approaches often combine restricting access to harmful content with strengthening coping skills and social connectedness, rather than relying on restriction alone (CDC suicide prevention resource).

Recovery Content Caught in the Crossfire

Perhaps the most striking finding involves recovery narratives, posts where people share their healing journey, visible scars, or personal progress. More than a fifth of respondents had a post removed or censored specifically because it showed self-harm scars, and the overwhelming majority found that experience harmful rather than protective.

This is where automated moderation systems tend to struggle. An algorithm trained to flag self-harm imagery often can’t distinguish between:

• A photo glorifying or instructing self-harm
• A recovery story showing healed scars as a symbol of resilience
• Educational or awareness content meant to reduce stigma
• A cry for help that needs a supportive response, not removal

Interviewees in the study described feeling punished for being open about their history, which is the opposite of what platforms presumably intend. If someone shares a recovery story and gets it taken down, the message they receive is that their experience is shameful or unwelcome, not that the platform is trying to keep others safe.

What Users Actually Want: More Control, Not More Restriction

The clearest message from the data is this: 87.8% of respondents wanted more personal control over what they see, not necessarily less content overall. That’s an important distinction for platform designers and policymakers to sit with.

Control-based tools that came up as desirable in the research included:

• Customizable content filters that let users opt in or out of specific topics rather than platforms deciding universally
• Clear, specific content warnings rather than vague or generic ones
• Adjustable sensitivity settings tied to a person’s own mental health needs at a given time
• Age-appropriate defaults for younger users, with more limited features until they reach a certain age or verification threshold
• Easy reporting tools that don’t require navigating confusing menus

Dr. Marchant emphasized that these controls “need to be easy to use and intuitive” if they’re going to meaningfully increase a young person’s agency online. A setting buried five menus deep doesn’t help anyone in a moment of crisis.

Building Genuinely Safe Spaces for Connection

The study doesn’t stop at criticizing what’s not working. It also makes a case for what social media platforms could be doing better. Beyond restriction, there’s an opportunity to use these platforms as tools for:

Signposting to Support

Instead of simply removing a self-harm-related post, platforms can attach it to crisis resources, helplines, or mental health services. This turns a moment of risk into a potential entry point for help, rather than a dead end.

Reducing Stigma Through Visibility

Self-harm remains heavily stigmatized, and hiding all evidence of it online doesn’t change that. Thoughtfully moderated recovery content can normalize seeking help and show people that healing is possible, which is a message that’s hard to deliver through silence.

Supporting Peer Communities

For many people, especially those who don’t have access to formal mental health services, online peer support is not a substitute for care. It’s often the first and most accessible form of care they encounter. The study’s authors argue platforms should treat these communities as an asset to strengthen, not a liability to eliminate.

The Samaritans, a UK-based charity focused on suicide prevention, has echoed similar demands in its own research, calling on platforms to give users more control over what they see, ensure self-harm and suicide content is never proactively pushed to users, and improve the support resources available directly within the platform experience (Samaritans research on social media and self-harm). That alignment between independent charity research and university-led academic study adds real weight to the case for co-designed, evidence-based policy rather than reactive bans.

What This Means for Parents, Platforms, and Policymakers

For parents and caregivers, the takeaway isn’t that social media is inherently safe or inherently dangerous. It’s that open conversation tends to work better than surveillance or outright bans. If a young person feels they’ll lose access entirely for admitting they’ve seen troubling content, they’re less likely to bring it up at all.

For platforms, the research is a fairly direct call to action: invest in moderation systems that can tell the difference between harmful and supportive content, build content controls that are actually easy to find and use, and stop treating every self-harm-related post as equally dangerous regardless of context.

For policymakers, the message from both researchers is consistent: any new restriction or safety feature should be co-designed with young people and those with lived experience, then tested thoroughly before wide rollout. Professor John was direct about this, noting that online interventions should be “co-designed, tested and evaluated before mass roll-out” rather than implemented first and studied later.

Practical Steps Platforms Could Take Right Now

Based on the study’s findings, a few concrete changes stand out as achievable in the near term:

1. Separate recovery content from harmful content in moderation guidelines, with clear criteria moderators can actually apply consistently.
2. Replace generic warnings with specific ones so users can make informed decisions about what they engage with.
3. Give users adjustable content controls instead of applying the same restriction level to everyone.
4. Build in direct links to crisis support wherever self-harm-related content appears, rather than simply deleting it.
5. Create simplified, age-tiered experiences for younger users without removing them from the platform entirely.
6. Regularly evaluate the real-world impact of any restriction, using feedback from affected users rather than assuming the policy is working.

A Note on Getting Support

If you or someone you know is struggling with self-harm or suicidal thoughts, please know that support is available and reaching out matters. In the US, you can call or text 988 to reach the Suicide & Crisis Lifeline. In the UK, Samaritans can be reached free at 116 123. These conversations are hard, but you don’t have to have them alone.

Conclusion

The Swansea University study makes one thing clear: the relationship between social media and self-harm is far more nuanced than “ban it” or “leave it alone.” Restrictions introduced since 2019 have made a real difference for some users, but they’ve also inadvertently silenced recovery stories, increased isolation, and pushed some young people toward less visible, less safe corners of the internet. What people with lived experience are actually asking for isn’t less access. It’s more control, clearer information, and spaces that treat their recovery as something to support rather than something to hide. Getting there will take platforms, researchers, and policymakers doing the harder work of building and testing solutions alongside the people those solutions are meant to protect, rather than deciding for them from the outside.