# PAST2HARM: A Simple Adaptive Past Tense Attack for Jailbreaking Multimodal AI

arXiv: https://arxiv.org/abs/2605.27545 · HTML: https://mlsanity.com/paper/2605.27545/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Vision-language and text-to-image safety, one of 12 directions in LLM safety and security.
- Builds on 20 of its 31 references that are in the map.
- First posted on arXiv 26 May 2026.

## Placed in

- LLM safety and security / [Vision-language and text-to-image safety](https://mlsanity.com/safety/vision-language-and-text-to-image-safety/)

## Builds on (most cited first, 20 of 20)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis](https://mlsanity.com/paper/2307.01952/) (2023)
- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [Constitutional AI: Harmlessness from AI Feedback](https://mlsanity.com/paper/2212.08073/) (2022)
- [Jailbroken: How Does LLM Safety Training Fail?](https://mlsanity.com/paper/2307.02483/) (2023)
- [HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal](https://mlsanity.com/paper/2402.04249/) (2024)
- [Jailbreaking Black Box Large Language Models in Twenty Queries](https://mlsanity.com/paper/2310.08419/) (2023)
- [Tree of Attacks: Jailbreaking Black-Box LLMs Automatically](https://mlsanity.com/paper/2312.02119/) (2023)
- [Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks](https://mlsanity.com/paper/2404.02151/) (2024)
- [FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Prompts](https://mlsanity.com/paper/2311.05608/) (2023)
- [Unified Concept Editing in Diffusion Models](https://mlsanity.com/paper/2308.14761/) (2023)
- [Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Models](https://mlsanity.com/paper/2307.14539/) (2023)
- [Red-Teaming the Stable Diffusion Safety Filter](https://mlsanity.com/paper/2210.04610/) (2022)
- [Ring-A-Bell! How Reliable are Concept Removal Methods for Diffusion Models?](https://mlsanity.com/paper/2310.10012/) (2023)
- [MMA-Diffusion: MultiModal Attack on Diffusion Models](https://mlsanity.com/paper/2311.17516/) (2023)
- [Circumventing Concept Erasure Methods For Text-to-Image Generative Models](https://mlsanity.com/paper/2308.01508/) (2023)
- [Does Refusal Training in LLMs Generalize to the Past Tense?](https://mlsanity.com/paper/2407.11969/) (2024)
- [Cannot or Should Not? Automatic Analysis of Refusal Composition in IFT/RLHF Datasets and Refusal Behavior of Black-Box LLMs](https://mlsanity.com/paper/2412.16974/) (2024)
- [Universally Unfiltered and Unseen:Input-Agnostic Multimodal Jailbreaks against Text-to-Image Model Safeguards](https://mlsanity.com/paper/2508.05658/) (2025)

## Abstract (arXiv)

Jailbreak attacks on multimodal AI systems remain underexplored, even though unsafe image generation can have more severe consequences than unsafe text and current defenses are relatively immature. We introduce PAST2HARM, a simple yet effective adaptive jailbreak framework that bypasses refusal training in state of the art multimodal text to image models. Building on prior findings that past tense reformulations can evade safeguards, PAST2HARM systematically exploits this vulnerability in multimodal generative AI. We characterize the attack along two dimensions. First, breadth: through temporal deepening, the framework incrementally strengthens historical anchoring and archival cues, eroding refusal boundaries across models with varying alignment strength. Second, depth: via iterative escalation after initial compliance, we probe the upper bound of harmful generation, measuring severity using a scalar severity jailbreak metric evaluated by a language model acting as a judge. We find that mid conversation turns form peak vulnerability windows, where harmfulness increases before plateauing and eventually undergoing semantic inversion. We evaluate PAST2HARM on three models Gemini Nano Banana Pro, GPT Image 2, and SD XL achieving attack success rates of 83 percent, 67 percent, and 100 percent in a black box, gradient free setting. Adversarial prompts also transfer across models, with cross model success rates above 50 percent. The attack elicits diverse harmful outputs, including explicit sexual content, political disinformation, historical denial narratives, hate speech, and self harm glorification. We further release a curated benchmark of prompts, reformulations, and outputs as a resource for red teaming and alignment. Our results expose fundamental brittleness in current safeguards and highlight the need for stronger multimodal safety training.