Whenever the US tries to make sense of a high-profile mass shooting, it inevitably turns to one source: the social media accounts of the suspect. Law enforcement, reporters and the public scrutinize these digital footprints, hoping to find clues about a possible motive.

Less explored, however, is the role social media platforms like Instagram and YouTube play in shootings that happen in underserved Black and Latino communities and are scarcely covered outside of local crime news. These shootings, Thomas Abt, the lead author of a new Violence Reduction Center white paper on the topic told the Guardian, are increasingly being fueled by online disputes and barbs being traded back and forth in songs and music videos and shared online.

This dynamic has been acknowledged anecdotally by law enforcement, those who work to stop shootings in the neighborhoods where they’re concentrated and the young people who see the disputes go from online to the real world, but little reporting and peer-reviewed research has been done on the topic, leading to a dearth of tailored solutions for violence prevention workers and law enforcement to try, Abt and his co-authors argue.

In an interview with the Guardian, Abt describes what he and his co-authors found during the symposium their report was based on the roles that violence prevention workers and social media companies can play to identify and address escalating interactions online.

This interview has been edited and condensed for clarity.

If you work in this field – or any – over the past five to 10 years, you have witnessed social media playing an increasingly large role in everyday life. And in this context, that means a lot of interactions that were happening in the real world are now happening online.

This white paper was born out of a sense of frustration. In the violence reduction field and among law enforcement, this is no secret. There’s a broad awareness that social media is playing an increasingly larger role. But the response is very anecdotal and ad hoc, so this was the VRC’s attempt to say: “Hey, we need to have a more structured conversation. We need to start planning, implementing, testing and evaluating. We just need to really start addressing this, not just amongst ourselves, we also need to really start calling for accountability and partnership from the platforms on which a lot of this communication is happening.”

How have people working on the ground seen and responded to social media content’s role in real world violence?

I think the thing that’s challenging for violence prevention groups is they understand that they need to have one foot in the social media world and one foot in the actual physical community world. And they’re doing that, but there’s not a systematic way of aggregating folks’ experience to develop best practices.

We had this really fantastic youth panel that was organized by Antonio Moore, and he brought in youth from around the country who are fluent in this stuff. And one of the things that we came to understand is that there is a really toxic dynamic going on where talking about violence and promoting violence and discussing violence in great detail and great specificity gets you more attention. And that attention, as we all know now, can be monetized.

And so it’s a really unhealthy dynamic that we need to address, and in a sense, folks from outside the community are egging on this violence inside these communities.

It’s a new version of an old dynamic. Mass shootings are so shocking that it gets a lot of media attention and the everyday community violence which is not really getting the attention it deserves, even though it causes many more deaths per year.

So, what needs to be done to turn this dynamic on its head?

I think any stakeholder that wants to be effective in this space need to have a structured conversation about what the best practices are and how we can develop evidence, study models and test them and do these things that we’ve done around mass shootings. That means law enforcement, community-based organizations, funders and service providers.

I was quite disappointed in the fact that we invited every major social media platform to participate in this symposium. And while we had some conversations via email and had some meetings offline, we couldn’t get a formal representative to participate in our panels and talk about that perspective. And that needs to change. They need to come to the table and start talking about this and start problem-solving.

I also think one of the challenges here is that social media can be something of a black box. If you’re outside these platforms, you don’t know what you don’t know. And so we need them to help us understand how these algorithms work and what they’re prioritizing. What are the ways that we can be most efficient in terms of monitoring content and taking down harmful content?

It’s incumbent on us to put pressure on them to do so because, you know, they are enormously important. And I do think that there’s a role for accountability and asking them to do better in terms of monitoring content.

We need to do a few things like investing in local community-led digital expertise. A lot of these online communications are very context dependent. And if you don’t know the context, you might not be able to distinguish what is truly a threat of violence from what isn’t.

A lot of people at the symposium expressed concerns with over-surveillance and over-enforcement, you know, sort of a digital stop and frisk, so we have to bring people to the table – often gen Z and gen Alpha, who have the lived cultural experience, that will help address that. We need ethical, non-punitive digital early warning systems, and those need to be built outside the platforms and inside the platforms. We need to develop national best practices and training standards.