Self-harm images and videos are thriving on Elon Musk's Twitter, which is less capable than ever of dealing with it

• Insider identified a slew of content glorifying self-harm on Twitter, despite a blanket ban.
• Experts worry that Elon Musk's moves to roll back moderation are exacerbating a longstanding issue.

Graphic self-harm content is thriving on Twitter in the era of Elon Musk, an Insider investigation found.

Banned content is widespread, despite clear policies and a claim from Musk that he made child safety the site's "top priority" on taking over.

Anything glorifying self-harm is notionally banned from Twitter — but Insider easily found examples of every banned category of content. Outside researchers shared further examples showing the scale of the issue.

The problem of such content spreading is not unique to Musk's tenure, which began in late October. But experts have voiced concern that his moves to roll back moderation on the platform have left Twitter less able to react.

In December, Insider found dozens of photos and videos of self-injury, some including gory imagery of profusely bleeding wounds. Insider is not reproducing the imagery.

Some posts were widespread, with likes from more than 500 other accounts. Self-harm hashtags were also full of jokes and memes making positive references to the practice.

Insider made repeated requests for comment on this story, including to Musk directly, but did not receive a response. Musk closed Twitter's communications department when he took over.

A dark, persistent problem

The self-harm problem on Twitter has long been clear, predating Musk's takeover in late October.

In August, the Network Contagion Research Institute released a report highlighting an expanding community dedicated to self-harm videos and images. Visible posts included photos of bloody wounds, razor blades, and other forms of self-injury.

It said that in the space of a year, the use of several popular self-harm hashtags on Twitter had increased by 500%.

That report built on findings from the children's digital-rights group 5Rights Foundation in November 2021 that searches for "self-harm" generated that kind of imagery, rather than resources or help.

Musk has said he considers child-safety at Twitter to be inadequate, though focusing on sexual exploitation rather than self-harm.

On that topic, he said in an exchange on December 9 that Twitter had been negligent to the point of criminality by ignoring child exploitation on the platform. (Jack Dorsey, Twitter's founder and ex-CEO, called that assertion "false.")

Musk said that child safety had been made "top priority immediately," and has promoted details from Twitter executives of cracking down on that practice.

He added that Ella Irwin, who left Twitter at the start of Musk's tenure but then returned, had been appointed as head of trust and safety. Insider has reached out to Irwin for comment.

Self-harm, however, has not received the same attention. Insider could find no public statements about it from Musk-era leaders at Twitter, and the experts Insider spoke to had not noticed a change.

Clear policy, lax enforcement

Alex Goldenberg, lead intelligence analyst at the NCRI and corresponding author of its August 2022 report, told Insider that self-injury content can "act like a social contagion."

People posting about self-harm "react to the positive feedback of the community, much like anyone would react to the positive feedback in any social community, by engaging in more harmful behavior to receive more positive feedback," he said.

Goldenberg described platforms allowing the content to spread as "a recipe for disaster."

Twitter's suicide and self-harm policy, which Musk has not altered, says that posts "may not promote or encourage suicide or self-harm."

It includes explicit bans on the following:

• Self-harm imagery.
• Descriptions of self-inflicted injuries.
• Encouraging somebody to harm themselves.
• Seeking encouragement from others for self-harm.
• Sharing "information, strategies, methods or instructions" for self-harm.

Insider found examples of each of the above types of content on Twitter.

Goldenberg said other platforms handle self-harm content far better. He commended Reddit and Instagram for fostering communities that support people struggling with self-harm, and giving practical advice on how to seek help.

On Twitter, "it's the exact opposite," he said. "It's an accelerator for this behavior, which is why it's so dangerous."

Sebastian Scherr, a professor of digital health communication at the University of Augsburg, told Insider that it's long been a struggle to get social media platforms to commit to really investing in tackling self-harm content.

Scherr has published research that shows that banning self-harm imagery can be automated.

"It is really a low hanging fruit," he said. "I don't really understand why the big tech companies are not being more active on this."

Psychologist Pamela Paresky, an NCRI scholar who contributed to its 2022 report, shared some recent findings with Insider.

This included a screen-record from December 13 showing that self-harm posts were highly visible in searches for the number 988, the number of the National Suicide Prevention Lifeline.

She also found reams of content showing people cutting their arms and bodies.

As well as the screen recording, Paresky also showed Insider a video appearing to advertise razor blades, in which three young girls in school girl outfits self-harm and praise the quality of the blades.

Paresky told Insider she was horrified.

"I really cannot understand how so much self-harm content is on the platform," she said. "Where was anybody who was supposedly keeping people safe?"

Inconsistent even before Musk

Though much content goes unpoliced, some is aggressively enforced.

In October, Paresky was locked out of her account for a week and a half for posting screenshots of tweets she found under self-harm hashtags. She said it required ten appeals to be allowed back on the platform to continue posting anti-self-harm content.

This all happened before Musk took over Twitter, laying off much of its staff and issuing a slew of chaotic directives to remake the network.

Goldenberg and Paresky both said they hoped Musk would stand by past statements that child-safety is a priority. His behavior suggests that his attention is elsewhere.

Under Musk, enforcement was pared back sharply.

One of Musks' first acts as CEO was to implement mass layoffs, closing entire departments of Twitter at a stroke.

Among the cuts were contractors who were responsible for monitoring and combating toxic content on the platform.

Researchers from Montclair State University described an almost immediate spike in hateful content.

On December 12, Musk went further, and dissolved Twitter's Trust and Safety Council, the independent advisory group that advocated for safety on the site. Its remit included addressing self-harm content.

Daniel Romer, an expert who published a 2019 paper exploring links between self-harm content online and actual self-harm, told Insider he was not optimistic for the future.

He told Insider that Musk was clearly interested in banning accounts critical of him personally, but that he hasn't taken down "very much of anything else."

"I doubt seriously with Musk taking over Twitter that they're doing much of anything in regard to self-harm," he said.

Lauren Mak, a censorship and online safety expert at online safety information website VPNOverview, told Insider that the risk from self-harm content on Twitter "is only increasing," describing Musk-era moderation as "completely arbitrary."

Content moderation can't be the whole answer, she argued, but stressed it was vital to keep people safe.

"Content that glorifies self-harm has been able to thrive on Twitter, and will likely only get worse if things don't change soon," she said.