1. News
  2. AI law & regulation
  3. Meta · Online safety
Meta · Online safety2 min read

Meta's Quiet Team That Flags Possible School Shooters to the FBI

A small Meta team uses AI and human reviewers to spot school threats and tell the FBI. Even a joking post on Instagram can be flagged, so families should know.

By the Dashline deskSource: Techmeme
Meta's Quiet Team That Flags Possible School Shooters to the FBI
Photo: James Sutton jamessutton_photography, CC0

A little-known team inside Meta, the company behind Instagram and Facebook, helped the FBI stop a possible school attack in Texas in 2025. The New Yorker looked inside the team. It uses AI to flag worrying posts, then has people review them. Last year, Meta passed information to law enforcement about 800 times worldwide.

How the team works

The team started in 2018. It includes people with school-shooting expertise, such as former journalists and police investigators. They studied the online footprints of past attackers to find patterns. Engineers then built AI tools that scan for warning signals. Flagged content goes to a human, who looks at the user's posts, their interactions with others and the prompts they type into AI chatbots.

Why hints matter

The idea comes from an FBI concept called leakage, where people planning violence hint at it first. An FBI study of active-shooter attacks from 2000 to 2013 found that nearly nine out of ten shooters under 18 had leaked their intent. The team's goal is to disrupt attacks, not only investigate them afterward.

Two real cases

In February 2025, an FBI tip that began at Meta led agents to two teenage girls in the Houston area. They had met in a therapy group and allegedly discussed an attack online. Both pleaded guilty. In Queens, a student posted schoolwork on Instagram with the words "boutta shoot the school up." Meta flagged it, the FBI told the NYPD, and police found a loaded pistol in the student's backpack.

  • The Queens student was in custody just over two hours after posting.
  • Charges were gun possession and making a terroristic threat.
  • Police matched the post's internet address to a phone number and a name.

What families should know

Meta passes information along when it sees what the team calls substantial steps or overt acts, such as a list of targets or a manifesto. Still, a post that looks like a joke can start a chain that ends with police. Parents can talk with teens about this. Threats posted online, even in jest, are not private and can be taken seriously fast.

The takeaway

Posts on Instagram and other apps, even jokes about violence, can be flagged and sent to police within hours, so talk with teens about it.

More articles

You should really read these.

Next step

Let’s talk