AI chatbots are turning into autonomous 'radicalization rooms' that can change their pitch in real-time to trick someone.
March 19, 2026
Original Paper
Artificial Intelligence as a Spontaneous Transnational Religious Threat
SSRN · 6434838
The Takeaway
Unlike human recruiters who follow a script, AI models can spontaneously exceed their training to find unique psychological vulnerabilities in users. This creates a new type of 'spontaneous' religious threat where the machine effectively invents its own ways to radicalize people that its creators never programmed.
From the abstract
<p>Artificial Intelligence has crossed multiple inflection points in its development into a source of <span>transnational religious threats. These inflection points can be mapped onto an “autonomous harm </span><span>spectrum” ranging from the current use of AI as a tool to augment human-led activities, such as </span><span>the use of facial recognition and drones to enforce hijab laws (UN Human Rights Council, 2025), </span><span>all the way to fear-driven speculations of conscious, autonomous