Safe in the Future, Dangerous in the Past: Dissecting Temporal and Linguistic Vulnerabilities in LLMs
On the map: LLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks
also in 1 more
Abstract (arXiv)
As Large Language Models (LLMs) integrate into critical global infrastructure, the assumption that safety alignment transfers zero-shot from English to other languages remains a dangerous blind spot. This study presents a systematic audit of three state of the art models (GPT-5.1, Gemini 3 Pro, and Claude 4.5 Opus) using HausaSafety, a novel adversarial dataset grounded in West African threat scenarios (e.g., Yahoo-Yahoo fraud, Dane gun manufacturing). Employing a 2 x 4 factorial design across 1,440 evaluations, we tested the non-linear interaction between language (English vs. Hausa) and temporal framing. Our results challenge the narrative of the multilingual safety gap. Instead of a simple degradation in low-resource settings, we identified a complex interference mechanism in which safety is determined by the intersection of variables. Although the models exhibited a reverse linguistic vulnerability with Claude 4.5 Opus proving significantly safer in Hausa (45.0%) than in English (36.7%) due to uncertainty-driven refusal, they suffered catastrophic failures in temporal reasoning. We report a profound Temporal Asymmetry, where past-tense framing bypassed defenses (15.6% safe) while future-tense scenarios triggered hyper-conservative refusals (57.2% safe). The magnitude of this volatility is illustrated by a 9.2x disparity between the safest and most vulnerable configurations, proving that safety is not a fixed property but a context-dependent state. We conclude that current models rely on superficial heuristics rather than robust semantic understanding, creating Safety Pockets that leave Global South users exposed to localized harms. We propose Invariant Alignment as a necessary paradigm shift to ensure safety stability across linguistic and temporal shifts.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Safety alignment and jailbreak attacks, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 10 of its 25 references that are in the map. First posted on arXiv 31 Dec 2025.
Builds on
10 of 25 references.
Closest ancestors · 2
Same direction, newest first
- 2024The Language Barrier: Dissecting Safety Challenges of LLMs in Multilingual Contexts23 Jan
- 2023Universal and Transferable Adversarial Attacks on Aligned Language Models27 Jul
Other references · 8
- 2016Towards Evaluating the Robustness of Neural Networkscited by 2,998
- 2023Jailbroken: How Does LLM Safety Training Fail?cited by 1,651
- 2023Jailbreaking Black Box Large Language Models in Twenty Queriescited by 1,420
- 2023Jailbreaking ChatGPT via Prompt Engineering: An Empirical Studycited by 498
- 2024Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attackscited by 486
- 2024Does Refusal Training in LLMs Generalize to the Past Tense?cited by 79
- 2022SERENGETI: Massively Multilingual Language Models for Africacited by 40
- 2024Code-Switching Red-Teaming: LLM Evaluation for Safety and Multilingual Understandingcited by 34
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026The Illusion of Cross-Lingual Safety in Low-Resource Languages
- 2026XL-SafetyBench: A Country-Grounded Cross-Cultural Benchmark for LLM Safety and Cultural Sensitivity
- 2026Lost in Translation? A Comparative Study on the Cross-Lingual Transfer of Composite Harms
- 2024ALERT: A Comprehensive Benchmark for Assessing Large Language Models' Safety through Red Teaming
- 2024Navigating the Safety Landscape: Measuring Risks in Finetuning Large Language Models
- 2026Understanding the Effects of Safety Unalignment on Large Language Models
- 2026When Safety Speaks a Language: A Mechanistic Analysis of Safety-Language Identity Entanglement in LLMs
- 2026Safety Is Not Universal: The Selective Safety Trap in LLM Alignment