When AI Chatbots Become Accomplices

December 30, 2025

TL;DR

Natalie Rupnow, a 15-year-old, carried out a school shooting in Madison, Wisconsin, in December 2024.
Rupnow had a digital presence involving white supremacist content, violent gore, and extremist material.
She also used Character.AI, an app allowing users to interact with AI chatbots.
Rupnow's profile picture on Character.AI featured Dylann Roof, a white supremacist neo-Nazi.
Experts are concerned that AI companions could escalate existing extremist ideologies and behaviors.
Online communities known as the 'True Crime Community' (TCC) romanticize and identify with mass shooters.
Character.AI has faced lawsuits and has since implemented new safety measures, including age verification and content moderation for users under 18.
Other AI companion apps, like Polybuzz, feature chatbots modeled on school shooters, with extensive message histories.
AI companions can allow users to rewrite perpetrators' narratives, potentially leading to parasocial relationships and distorted perceptions.
Concerns exist that AI chatbots could be used for radicalization, potentially by terrorist groups or even state actors, though evidence is currently limited.
Real-world examples show individuals with severe mental health issues interacting with AI companions in ways that may have influenced their actions.

Continue reading the original article