Social media rarely creates violence in isolation, but it can make existing conflict and harmful attitudes faster, more visible and harder to escape, according to a major new study commissioned by the Mayor of London’s Violence Reduction Unit.

“Social Media and Offline Violence” examined how online environments contribute to interpersonal violence affecting young people and to violence against women and girls in London. The work was conducted collaboratively by M&C Saatchi World Services, nisien.ai and Shout Out UK.

nisien.ai led the large-scale quantitative analysis. More than 260,000 public posts on X were collected and analysed over a 12-month period. Of these, 25,509 were confirmed as misogynistic and 10,978 as violence-related. The online data were examined alongside daily Metropolitan Police records covering the same period.

On days when misogynistic content formed a larger share of public conversation on X, police-recorded violence against women and girls also increased. The report treats this as a city-wide association rather than proof that individual posts directly caused particular offences.

The wider study also included 22 interviews and eight focus groups with young people, practitioners, teachers and youth workers. These accounts showed how online audiences can raise the stakes of disputes, how public humiliation can create pressure to respond, and how repeated exposure can normalise misogynistic attitudes or desensitise young people to violent material.

The report describes social media as both a “mirror and motor” of harm: reflecting tensions that already exist while accelerating and extending them through public visibility, sharing and recommendation systems.

Its recommendations include safer platform design, investment in trusted relationships and rapid-response capacity, peer-led approaches to interrupt escalation, stronger education and improved support for those affected by image-based abuse. Following the publication of the report, MOPAC committed £5 million to prevention activity.

The study continues a central strand of HateLab research: moving beyond simple counts of harmful content to understand when online behaviour becomes connected to consequences in the offline world.

Download the full report