Any improvements that are not entangled with a learning objective, i.e., pure pre-processing, usually arising from clues found in the node features and graph structure.

AI-generated text detectors have become essential tools for maintaining content authenticity, yet their robustness against evasion attacks remains questionable. We present PDFuzz, a novel attack that exploits the discrepancy between visual text layout and extraction order in PDF documents. Our method preserves exact textual content while manipulating character positioning to scramble extraction sequences. We evaluate this approach against the ArguGPT detector using a dataset of human and AI-generated text. Our results demonstrate complete evasion: detector performance drops from (93.6 $\pm$ 1.4) % accuracy and 0.938 $\pm$ 0.014 F1 score to random-level performance ((50.4 $\pm$ 3.2) % accuracy, 0.0 F1 score) while maintaining perfect visual fidelity. Our work reveals a vulnerability in current detection systems that is inherent to PDF document structures and underscores the need for implementing sturdy safeguards against such attacks. We make our code publicly available at https://github.com/ACMCMC/PDFuzz.

While chain-of-thought (CoT) monitoring is an appealing AI safety defense, recent work on "unfaithfulness" has cast doubt on its reliability. These findings highlight an important failure mode, particularly when CoT acts as a post-hoc rationalization in applications like auditing for bias. However, for the distinct problem of runtime monitoring to prevent severe harm, we argue the key property is not faithfulness but monitorability. To this end, we introduce a conceptual framework distinguishing CoT-as-rationalization from CoT-as-computation. We expect that certain classes of severe harm will require complex, multi-step reasoning that necessitates CoT-as-computation. Replicating the experimental setups of prior work, we increase the difficulty of the bad behavior to enforce this necessity condition; this forces the model to expose its reasoning, making it monitorable. We then present methodology guidelines to stress-test CoT monitoring against deliberate evasion. Applying these guidelines, we find that models can learn to obscure their intentions, but only when given significant help, such as detailed human-written strategies or iterative optimization against the monitor. We conclude that, while not infallible, CoT monitoring offers a substantial layer of defense that requires active protection and continued stress-testing.

This paper presents a holistic approach to attacker preference modeling from system-level audit logs using inverse reinforcement learning (IRL). Adversary modeling is an important capability in cybersecurity that lets defenders characterize behaviors of potential attackers, which enables attribution to known cyber adversary groups. Existing approaches rely on documenting an ever-evolving set of attacker tools and techniques to track known threat actors. Although attacks evolve constantly, attacker behavioral preferences are intrinsic and less volatile. Our approach learns the behavioral preferences of cyber adversaries from forensics data on their tools and techniques. We model the attacker as an expert decision-making agent with unknown behavioral preferences situated in a computer host. We leverage attack provenance graphs of audit logs to derive a state-action trajectory of the attack. We test our approach on open datasets of audit logs containing real attack data. Our results demonstrate for the first time that low-level forensics data can automatically reveal an adversary's subjective preferences, which serves as an additional dimension to modeling and documenting cyber adversaries. Attackers' preferences tend to be invariant despite their different tools and indicate predispositions that are inherent to the attacker. As such, these inferred preferences can potentially serve as unique behavioral signatures of attackers and improve threat attribution.

In the age of powerful AI-generated text, automatic detectors have emerged to identify machine-written content. This poses a threat to author privacy and freedom, as text authored with AI assistance may be unfairly flagged. We propose AuthorMist, a novel reinforcement learning-based system to transform AI-generated text into human-like writing. AuthorMist leverages a 3-billion-parameter language model as a backbone, fine-tuned with Group Relative Policy Optimization (GPRO) to paraphrase text in a way that evades AI detectors. Our framework establishes a generic approach where external detector APIs (GPTZero, WinstonAI, Originality.ai, etc.) serve as reward functions within the reinforcement learning loop, enabling the model to systematically learn outputs that these detectors are less likely to classify as AI-generated. This API-as-reward methodology can be applied broadly to optimize text against any detector with an accessible interface. Experiments on multiple datasets and detectors demonstrate that AuthorMist effectively reduces the detectability of AI-generated text while preserving the original meaning. Our evaluation shows attack success rates ranging from 78.6% to 96.2% against individual detectors, significantly outperforming baseline paraphrasing methods. AuthorMist maintains high semantic similarity (above 0.94) with the original text while successfully evading detection. These results highlight limitations in current AI text detection technologies and raise questions about the sustainability of the detection-evasion arms race.

Image watermarks have been considered a promising technique to help detect AI-generated content, which can be used to protect copyright or prevent fake image abuse. In this work, we present a black-box method for removing invisible image watermarks, without the need of any dataset of watermarked images or any knowledge about the watermark system. Our approach is simple to implement: given a single watermarked image, we regress it by deep image prior (DIP). We show that from the intermediate steps of DIP one can reliably find an evasion image that can remove invisible watermarks while preserving high image quality. Due to its unique working mechanism and practical effectiveness, we advocate including DIP as a baseline invasion method for benchmarking the robustness of watermarking systems. Finally, by showing the limited ability of DIP and other existing black-box methods in evading training-based visible watermarks, we discuss the positive implications on the practical use of training-based visible watermarks to prevent misinformation abuse.

Equivocation and ambiguity in public speech are well-studied discourse phenomena, especially in political science and analysis of political interviews. Inspired by the well-grounded theory on equivocation, we aim to resolve the closely related problem of response clarity in questions extracted from political interviews, leveraging the capabilities of Large Language Models (LLMs) and human expertise. To this end, we introduce a novel taxonomy that frames the task of detecting and classifying response clarity and a corresponding clarity classification dataset which consists of question-answer (QA) pairs drawn from political interviews and annotated accordingly. Our proposed two-level taxonomy addresses the clarity of a response in terms of the information provided for a given question (high-level) and also provides a fine-grained taxonomy of evasion techniques that relate to unclear, ambiguous responses (lower-level). We combine ChatGPT and human annotators to collect, validate and annotate discrete QA pairs from political interviews, to be used for our newly introduced response clarity task. We provide a detailed analysis and conduct several experiments with different model architectures, sizes and adaptation methods to gain insights and establish new baselines over the proposed dataset and task.

In this work, we investigate the potential threat of adversarial examples to the security of face recognition systems. Although previous research has explored the adversarial risk to individual components of FRSs, our study presents an initial exploration of an adversary simultaneously fooling multiple components: the face detector and feature extractor in an FRS pipeline. We propose three multi-objective attacks on FRSs and demonstrate their effectiveness through a preliminary experimental analysis on a target system. Our attacks achieved up to 100% Attack Success Rates against both the face detector and feature extractor and were able to manipulate the face detection probability by up to 50% depending on the adversarial objective. This research identifies and examines novel attack vectors against FRSs and suggests possible ways to augment the robustness by leveraging the attack vector's knowledge during training of an FRS's components.

Abstract: Chatbots, or bots for short, are multi-modal collaborative assistants that can help people complete useful tasks. Usually, when chatbots are referenced in connection with elections, they often draw negative reactions due to the fear of mis-information and hacking. Instead, in this paper, we explore how chatbots may be used to promote voter participation in vulnerable segments of society like senior citizens and first-time voters. In particular, we build a system that amplifies official information while personalizing it to users' unique needs transparently. We discuss its design, build prototypes with frequently asked questions (FAQ) election information for two US states that are low on an ease-of-voting scale, and report on its initial evaluation in a focus group. Our approach can be a win-win for voters, election agencies trying to fulfill their mandate and democracy at large.

Botnet detectors based on machine learning are potential targets for adversarial evasion attacks. Several research works employ adversarial training with samples generated from generative adversarial nets (GANs) to make the botnet detectors adept at recognising adversarial evasions. However, the synthetic evasions may not follow the original semantics of the input samples. This paper proposes a novel GAN model leveraged with deep reinforcement learning (DRL) to explore semantic aware samples and simultaneously harden its detection. A DRL agent is used to attack the discriminator of the GAN that acts as a botnet detector. The discriminator is trained on the crafted perturbations by the agent during the GAN training, which helps the GAN generator converge earlier than the case without DRL. We name this model RELEVAGAN, i.e. ["relive a GAN" or deep REinforcement Learning-based Evasion Generative Adversarial Network] because, with the help of DRL, it minimises the GAN's job by letting its generator explore the evasion samples within the semantic limits. During the GAN training, the attacks are conducted to adjust the discriminator weights for learning crafted perturbations by the agent. RELEVAGAN does not require adversarial training for the ML classifiers since it can act as an adversarial semantic-aware botnet detection model. Code will be available at https://github.com/rhr407/RELEVAGAN.