Technology

lol, lets use AI to get an opinion of how good AI is at reducing shootings

what direction is the trend going in school shootings in the united states, more or less frequent over time?

The long-term trend in the United States is that school shootings have become more frequent overall, especially since the late 2010s and during/after the COVID-era years. Multiple major datasets and studies show a substantial upward trend over the past 25 years.

A few important nuances:

The exact numbers vary depending on the definition used: Some databases count any gunfire on school property (including after-hours incidents, accidental discharges, gang-related events, etc.). Others count only targeted attacks or mass-casualty events. Even with different definitions, most datasets still show an increase over time.

Some notable findings from recent research:

A study in the journal American Academy of Pediatrics found that the 2017–2022 school years had dramatically more school shootings than the previous two decades, with 2021–2022 reaching the highest level in their dataset.

KFF reported that student exposure to school shootings has nearly tripled since 1999, with the sharpest increases occurring during the pandemic years.

The broad K-12 School Shooting Database showed a peak around 2023, though 2024 and 2025 appear to have declined somewhat from that peak.

So the short version is:

Over decades: upward trend (more frequent). Very recently: there are signs of a modest decline from the post-pandemic peak, but rates remain historically high compared with the 1990s or early 2000s. how likely is it that the use in AI is reducing school shootings?

There’s currently no strong evidence that AI is already reducing school shootings in a measurable national way. The honest answer is that researchers are still in the “possible tool, unproven impact” stage.

What does seem plausible is that AI may help in a few narrow areas:

identifying threatening posts or communications, assisting school threat-assessment teams, sorting large amounts of tips or reports, flagging patterns humans might miss.

Researchers studying threat assessment think technology-assisted systems could potentially help identify credible threats earlier.

There’s also evidence that many attackers show warning signs beforehand — threats, writings, online posts, or concerning behavior — which is exactly the sort of pattern-detection task AI is often good at. One review noted that roughly 75–80% of school shooters communicated intent before attacks.

But there are major caveats:

Experts repeatedly say school shootings are rare and difficult to predict accurately. False positives are a huge problem — you do not want systems incorrectly flagging ordinary students. Many AI school-safety technologies (facial recognition, social-media scanning, behavior monitoring) have little rigorous evidence proving they actually prevent shootings.

Some researchers warn AI systems may introduce bias or over-policing concerns.

And there’s another side to this:

AI may also be creating new risks. Recent investigations and studies suggest some chatbots can, under certain circumstances, provide harmful or violent planning information despite safety guardrails.

So overall, the current evidence points to:

Potential modest preventive benefit through threat assessment and monitoring, but no demonstrated large-scale reduction in school shootings yet, and possibly some offsetting harms from unsafe AI behavior.

If you want, I can also break down:

which anti-shooting methods currently have the strongest evidence, whether armed guards actually reduce shootings, or how “threat assessment” programs statistically perform compared to other approaches.