Results for “Sam”

arXiv (research papers) Or Zion Eliav SI

Know Your Agent: Reconnaissance-Driven Pentesting of AI Agents

Traditional pentesting uses reconnaissance at each step to uncover unseen weaknesses, build stronger attacks, and advance the objective; we argue that AI agents require the same treatment. We formalize agent reconnaissance by modeling the process and identifying the knowledge assets it seeks to extract: what they are, how they are used, and which agent weaknesses they exploit to give adversaries leverage in indirect prompt injection attacks. We instantiate these insights in…

Read at arXiv (research papers)

Journal articles (OpenAlex) Qasir Shah · Journal of Philosophy of Education Geopolitics

When inclusion excludes: liberalism, Confucianism, and the Parkfield Community School protests

Abstract This article examines the 2019 Parkfield Community School protests in Birmingham as a case-based philosophical analysis in the moral and relational limits of liberal inclusion within public education. It critically engages Abeera Khan’s analysis of securitized liberalism, which shows how Muslim parental dissent was pathologized and rendered unintelligible within dominant civic discourses. At the same time, it draws on Nora Hämäläinen’s model of moral renegotiation…

Read at Journal articles (OpenAlex)

arXiv (research papers) Ruksat Khan Shayoni SI

NetInjectBench: Benchmarking Indirect Prompt Injection in Tool-Using Large Language Model Agents for Network Operations

Tool-using large language model (LLM) agents are attractive for network operations, but tickets, alerts, logs, runbooks, and ChatOps messages can carry indirect prompt injections. We present NetInjectBench, a 130-scenario benchmark that separates untrusted artifact text, trusted policy metadata, and evaluation labels for network-operation tool use. The sample contains 40 benign, 40 weak-attack, 40 strong-attack, and 10 approved high-impact change scenarios; each is evaluated…

Read at arXiv (research papers)

Rappler Rappler.com South China Sea

PH SolGen on 2016 Arbitral Award: ‘Enduring protection in a Republic of laws’

'We claimed our rights and in the same breath, honored everyone else's,' says the Philippine government's top lawyer

Read at Rappler

arXiv (research papers) Mohammad Dabaja SI

Promptable Concept Segmentation from Above: Evaluating SAM 3's Zero-Shot and One-Shot Capabilities in Remote Sensing

The deployment of large-scale foundation models, such as the Segment Anything Model 3 (SAM 3), promises a transition toward open-vocabulary, training-free computer vision. However, their capacity to generalize out-of-distribution to the complex, top-down geometric structures of Earth Observation imagery remains largely unquantified. Driven by SAM 3's performance disparities in highly specialized domains, we present a comprehensive, multi-task empirical evaluation across…

Read at arXiv (research papers)

Latinoamérica21 Orlando J. Pérez Geopolitics

Washington repeats in Cuba the same failed strategy of 60 years ago

Washington is once again betting on a strategy that already failed more than six decades ago: using economic sanctions to force political change in Cuba. Yet, while it seeks to isolate the regime, it also hampers reforms and pushes Havana closer to China. O post Washington repeats in Cuba the same failed strategy of 60 years ago apareceu primeiro em Latinoamérica 21.

Read at Latinoamérica21

arXiv (research papers) Elad David SI

Latent Undertow: How Ordinary Typos Break Probes

LLMs handle ordinary typing variation fluently: a typo or missing punctuation leaves both user intent and the model's response substantively unchanged. Yet probes that detect malicious prompts by reading the model's hidden states tell a different story: the same edit rotates the readout vector by 43--56 at the perturbed token, decaying below 15% within ~10 downstream tokens. Stacking ~3 common typos per message cuts a single-position prompt-injection probe's TPR@FPR$=1% by…

Read at arXiv (research papers)

Berkeley AI Research blog SI

Intelligence is Free, Now What? Data Systems for, of, and by Agents

... government of the people, by the people, for the people ... — Abraham Lincoln, Gettysburg Address (1863) The cost of AI is dropping rapidly. GPT-4-class capabilities cost roughly $30 per million tokens in early 2023; today the same runs under $1, and some providers are pushing costs below $0.10. Across benchmarks, inference prices have fallen between 9x and 900x per year, with a median decline near 50x. Even frontier models are getting dramatically cheaper each…

Read at Berkeley AI Research blog

Journal articles (OpenAlex) Rodrigo Antonio Landabur, Carlos Escobar-Campusano, Crhistian Rojo et al. · Behavioral Sciences Geopolitics

Psychometric Evaluation of the Identification with the Country Scale in a Chilean Sample

Group identification refers to a psychological connection with a group whereby individuals incorporate group-defining characteristics into their self-concept. The scale developed by Mael & Ashforth is one of the most used, but for knowledge, it has not been studied in Chile. This study analyzed the psychometric properties of this scale in a Chilean sample using the country as the reference group. We expected a one-factor structure and convergent validity. In a…

Read at Journal articles (OpenAlex)

arXiv (research papers) Zunchen Huang SI

Analyzing the Narration Gap in LLM-Solver Loops

Formal tools such as SAT and SMT solvers are increasingly embedded in language model reasoning pipelines when a safety or security critical question can be formulated in logic. Unlike chain of thought whose steps are sampled from the model distribution without formal guarantee, a solver produces a sound and independently verifiable answer. However, the soundness guarantee can be lost in the interaction between the solver and the model. The hybrid pipeline has three…

Read at arXiv (research papers)

arXiv (research papers) David Huang SI

Rapid Poison: Practical Poisoning Attacks Against the Rapid Response Framework

The Rapid Response (RR) framework, deployed in production systems, including Anthropic's ASL-3 safeguards, continuously improves jailbreak-detection classifiers. When new jailbreaks emerge that bypass these classifiers, Rapid Response generates synthetic variants for training, helping the model generalize from the new attacks and quickly adapt. We reveal that prompt injection can infiltrate this pipeline to deliver poisoned samples into the classifier's training set…

Read at arXiv (research papers)

Journal articles (OpenAlex) Asal Kamani Fard, Mohammad Mahdi Paydar · BMC Public Health Geopolitics

Women’s walking behaviour and subjective well-being in the middle-income neighbourhoods of Santiago, Chile

BACKGROUND: Walking remains the primary mode of everyday mobility for women in Chile, particularly in metropolitan areas with low to medium income populations. However, a range of barriers including among others concerns related to perceived insecurity have recently emerged and made walking more difficult. At the same time, subjective well-being has become an important topic in urban research over the past decade. This study examines the associations between women's walking…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Nicol Altamirano, David Sirlopú, Joseph Aslan et al. · International Journal of Social Psychology Revista de Psicología Social Geopolitics

Intergroup Threat and Tolerance Towards Venezuelan Immigrants in Chile / Tolerancia y amenaza intergrupal hacia inmigrantes venezolanos en Chile

This study compares levels of intergroup threat and tolerance towards Venezuelan immigrants in Chile among university students and members of the general population, examining whether distinct latent profiles emerged in the full pooled sample. A total of 490 Chilean adults participated in the study, including 289 university students ( M = 21.69, SD = 3.08) and 201 members of the general population ( M = 41.88, SD = 15.77). Participants completed validated Chilean versions of…

Read at Journal articles (OpenAlex)

arXiv (research papers) Mudit Sinha SI

Hiding in Plain Floats: Steganographic Carriers for Indirect Prompt and Content Injection

Text-centered prompt-injection defenses assume that the malicious signal is visible in one of the inspected text views. We study a reproducible LLM01-style indirect prompt/content-injection failure mode where that assumption breaks: a payload caught in plain English slips past the same detector when it is transported as structured float parameters and reconstructed only as fragmented telemetry. Across 14,400 attacked real-model trials on three commercial LLM APIs from…

Read at arXiv (research papers)

arXiv (research papers) Kargi Chauhan SI

Caught in the Act(ivation): Toward Pre-Output and Multi-Turn Detection of Credential Exfiltration by LLM Agents

LLM agents often place sensitive credentials in the same context window as untrusted retrieved content, creating a direct path for indirect prompt injection to induce credential exfiltration. We study this failure mode through three complementary defenses. First, we ask whether activation probes can detect credential access before output tokens are emitted. Second, we construct honeytokens from format-specific character models and calibrate detection with split conformal…

Read at arXiv (research papers)

arXiv (research papers) Ryle Goehausen SI

Gate AI: LLM Security Benchmark Evaluation Methodology and Results

Published evaluations of prompt-injection and jailbreak detectors for Large Language Models often suffer from two systematic weaknesses: per-dataset threshold tuning and undisclosed operating points. We describe an evaluation harness that addresses both. The detector under evaluation is scored across 16 public benchmarks (12,111 samples) using 5-fold cross-validation. StratifiedKFold (by row) is the headline pass; a parallel StratifiedGroupKFold pass over a composite key…

Read at arXiv (research papers)

arXiv (research papers) Hiskias Dingeto SI

AgentRedBench: Dynamic Redteaming and Integration-Aware Defense for LLM Agents over SaaS Integrations

Indirect prompt injection in tool-use agents is a concrete production threat: LLM agents read from integrations (third-party services such as Gmail, Salesforce, or Jira accessed through tool calls) whose response content the user neither writes nor controls. Existing benchmarks under-measure the threat: most cover only a handful of integrations with the same attack payload replayed across runs, and open-source guards are trained on chat-style data rather than tool-response…

Read at arXiv (research papers)

The Guardian Belarus Phuong Le

The Guest review – the risks of protecting refugees in the Poland-Belarus danger zone

Zvika Gregory Portnoy and Zuzanna Solakiewicz’s documentary lays bare the problems faced by refugees and the compassion of good samaritans It all begins with a knock. In a small Polish town on the border with Belarus, Maciek and his family have taken in 27-year-old Alhyder, a Syrian refugee seeking shelter from the freezing weather and police patrols. Since 2021, the area has become increasingly militarised after Vladimir Putin and Alexander Lukashenko, in a purely political…

Read at The Guardian Belarus

Journal articles (OpenAlex) Wenjing Fu, Hailong Zhang, Zicheng Wang et al. · Radiocarbon South China Sea

Assessments of radiocarbon measurements of dissolved organic carbon for small/ultra-small samples by MICADAS

Abstract The radiocarbon ( 14 C) composition of dissolved organic carbon (DOC) provides critical constraints on its sources, transformation and fates. UV-irradiation is an effective approach to completely oxidize DOC to CO 2 for radiocarbon measurements. In this study, we aimed to evaluate the performance of radiocarbon measurements for small/ultra-small size natural DOC samples using mini carbon dating system (MICADAS) using CO 2 gas as the target. The procedure blank was…

Read at Journal articles (OpenAlex)

arXiv (research papers) Syed Nazmus Sakib SI

The Surface You Test Is Not the Surface That Breaks

Prompt-injection benchmarks for LLM agents typically test attacks through a single injection surface and report the resulting attack success rate as a property of the model. We ask whether those robustness conclusions remain stable when the same adversarial content enters through a different part of the agent interface. Using AgentDojo, we evaluate 13 LLMs across four task suites and place a byte-identical payload either in a tool output or in the tool description. This…

Read at arXiv (research papers)

Journal articles (OpenAlex) Rose McKeon Olson, Josephine Ndanemah, Daniel Nguyen et al. · Archives of Public Health Geopolitics

Barriers to healthcare and justice for sexual and gender-based violence survivors in Sierra Leone: A gender-integrated socio-ecological analysis

BACKGROUND: Sexual and gender-based violence (SGBV) is a threat to women and girls' health and justice globally, and is widely prevalent in Sierra Leone. This study identifies key barriers to high-quality healthcare and justice for SGBV survivors in Sierra Leone using a gender-integrated socio-ecological model. METHODS: In this qualitative study, purposive sampling was performed with maximum variation by sector and region. From December 2023 to February 2024, Sierra Leonean…

Read at Journal articles (OpenAlex)

arXiv (research papers) Ruoqi Guo SI

MIRAGE: Context-Aware Prompt Injection against Mobile GUI Agents via User-Generated Content

Mobile graphical user interface (GUI) agents driven by vision-language models (VLMs) perceive the screen as rendered pixels and choose actions from what they see, so they cannot reliably separate trusted interface elements from user-generated content. We present MIRAGE (Mobile Injection of Realistic Adversarial GUI Examples), a pipeline that turns benign mobile screenshots into prompt-injection samples by placing attacker-controlled text into ordinary user-generated content…

Read at arXiv (research papers)

Journal articles (OpenAlex) Simon Rushton · Contemporary Security Policy Geopolitics

Global health security: The retreat of the liberal international order and the prospects for cooperation in a multi-order world

This Forum article examines what the emergence of a multi-order world might mean for the future of international cooperation over global health security. It argues that recent years have seen a retreat of key states of the liberal international order from the internationalist vision of collective disease control, weakening its foundations. At the same time, leading states of other international orders, such as Russia and China, have been resistant to some of the core…

Read at Journal articles (OpenAlex)

arXiv (research papers) Oliver Mel Allen Geopolitics

Same View, Different Visions -- Visual Nature Content on Social Media

Digital nature values formed through online interactions with nature might incentivize new types of environmental stewardship, but these values may also be appropriated by those who seek to undermine sustainable transformation, such as right-wing populist movements and fossil fuel industries. Specific aesthetics and visual cultures may be appropriated, borrowing their credibility or influence, and layers of platform affordances, internet cultures, and norms add additional…

Read at arXiv (research papers)

Journal articles (OpenAlex) Yihang Sui, Jiayi Jin, Ayşe Özbil Torun · Land Geopolitics

Heritage Conservation as Degrowth Practice: Multi-Scalar Analysis of Gasholder Adaptive Reuse in London and Edinburgh

Industrial heritage adaptive reuse occupies a structurally privileged position for degrowth: heritage listing already institutionalises material sufficiency as a regulatory obligation, mandating low intervention and resisting the demolish-and-replace logic of resource-intensive development. Yet this regulatory floor imposes no ceiling on how protected structures are programmed or who benefits; the same statutory instrument can produce different schemes depending entirely on…

Read at Journal articles (OpenAlex)

arXiv (research papers) Igor Bogdanov SI

FORGE: Self-Evolving Agent Memory With No Weight Updates via Population Broadcast

Can LLM agents improve decision-making through self-generated memory without gradient updates? We propose FORGE (Failure-Optimized Reflective Graduation and Evolution), a staged, population-based protocol that evolves prompt-injected natural-language memory for hierarchical ReAct agents. FORGE wraps a Reflexion-style inner loop, where a dedicated reflection agent (using the same underlying LLM, no distillation from a stronger model) converts failed trajectories into reusable…

Read at arXiv (research papers)

arXiv (research papers) Jie Zhang SI

Laundering AI Authority with Adversarial Examples

Vision-language models (VLMs) are increasingly deployed as trusted authorities -- fact-checking images on social media, comparing products, and moderating content. Users implicitly trust that these systems perceive the same visual content as they do. We show that adversarial examples break this assumption, enabling \emph{AI authority laundering}: an attacker subtly perturbs an image so that the VLM produces confident and authoritative responses about the \emph{wrong} input…

Read at arXiv (research papers)

Journal articles (OpenAlex) Vladimir Miroljub Cvetković, Dalibor Milenković, Jasmina Bašić et al. · Safety Geopolitics

Predictive Model of Community Disaster Resilience Across Serbia: A BRIC–DROP Composite Index and Spatial Patterns

Community disaster resilience is increasingly guiding risk-reduction investments, but in many Southeast European settings, comparable subnational data remain scarce. This study assesses perceived community disaster resilience across Serbia by combining BRIC–DROP dimensions into a single index and analyzing differences across hazard types and sociodemographic factors. A cross-sectional household survey was conducted using multistage random sampling and the “next birthday”…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Alma Harrison · Journal of the Arkansas Academy of Science Geopolitics

‘Enemy Communist Cocksuckers’: Conceptualizing Patriotic Masculinity in Threats to Election Officials

Research on political violence demonstrates how certain aspects of masculinity are linked to asserting ideological authority and reinforcing exclusionary social identities through physical aggression, such as assaults, homicides, and gender-based violence. However, little research analyzes the use of violent threats as masculinity performances to achieve these same ends. To address this gap, this study analyzes messages from qualitative interviews with 26 election officials…

Read at Journal articles (OpenAlex)

arXiv (research papers) Tosin Adewumi SI

BatteryPass-12K: The First Dataset for the Novel Digital Battery Passport Conformance Task

We introduce a novel task of digital battery passport (DBP) conformance classification and introduce the first public benchmark for the task: BatteryPass-12K, created synthetically from real pilot samples. This is as the EU's battery regulation on DBPs comes into effect soon and there exists no public dataset. We evaluated 22 language models (LMs) in zero-shot inference, spanning small LMs (SLMs), mixture of experts (MoEs), and dense LLMs. We also conducted analysis…

Read at arXiv (research papers)

Older stories