[2603.01942] Ignore All Previous Instructions: Jailbreaking as a de-escalatory peace building practise to resist LLM social media bots

Computer Science > Human-Computer Interaction arXiv:2603.01942 (cs) [Submitted on 2 Mar 2026] Title:Ignore All Previous Instructions: Jailbreaking as a de-escalatory peace building practise to resist LLM social media bots Authors:Huw Day, Adrianna Jezierska, Jessica Woodgate View a PDF of the paper titled Ignore All Previous Instructions: Jailbreaking as a de-escalatory peace building practise to resist LLM social media bots, by Huw Day and 2 other authors View PDF HTML (experimental) Abstract:Large Language Models have intensified the scale and strategic manipulation of political discourse on social media, leading to conflict escalation. The existing literature largely focuses on platform-led moderation as a countermeasure. In this paper, we propose a user-centric view of "jailbreaking" as an emergent, non-violent de-escalation practice. Online users engage with suspected LLM-powered accounts to circumvent large language model safeguards, exposing automated behaviour and disrupting the circulation of misleading narratives. Comments: Subjects: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI) Cite as: arXiv:2603.01942 [cs.HC]   (or arXiv:2603.01942v1 [cs.HC] for this version)   https://doi.org/10.48550/arXiv.2603.01942 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Huw Day [view email] [v1] Mon, 2 Mar 2026 14:57:13 UTC (279 KB) Full-text links: Access Paper: View a PDF of the paper titled Ignore All Previ...