Skip to main content
This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .

Radicalisation

A Collaborative, Human-Centred Taxonomy of AI, Algorithmic, and Automation Harms

Abercrombie et al. (2024)

Sub-category
Risk Domain

Highly personalized AI-generated misinformation creating “filter bubbles” where individuals only see what matches their existing beliefs, undermining shared reality, weakening social cohesion and political processes.

"Radicalisation - Adoption of extreme political, social, or religious ideals and aspirations due to the nature or misuse of an algorithmic system, potentially resulting in abuse, violence, or terrorism."(p. 18)

Other risks from Abercrombie et al. (2024) (53)