Results for “Again”

The Guardian Syria William Christou in Beirut Geopolitics

Syrian court sentences Bashar al-Assad to death in absentia for war crimes

Verdict relates to atrocities committed during the country’s 14-year civil war in which about 500,000 people died A Syrian court has sentenced the country’s former president Bashar al-Assad and his brother Maher to death in absentia for crimes against humanity and war crimes during the 14-year civil war, in which about half a million people died. The two men fled to Moscow after being ousted by a lightning rebel offensive in December 2024 which toppled the dictatorial regime…

Read at The Guardian Syria

Journal articles (OpenAlex) Ricardo Rudas Meo · Asian Ethnicity Geopolitics

The frontier on the front lines: the Chinese Communist discourse on China’s southwest and its peoples during the Korean War, 1950–1953

This article argues that the Korean War (1950–1953) was a discursive event through which the Chinese Communist Party (CCP) sought to buttress its vision of Chinese nationhood and territoriality within the newly-founded People’s Republic of China. It examines the CCP’s political discourse on China’s ethnoculturally-diverse southwestern borderlands and so-called ‘minority nationalities’ (shaoshu minzu) against the Korean War background. Through an analysis of People’s Daily…

Read at Journal articles (OpenAlex)

arXiv (research papers) Sihan Hou SI

Toward Metacognitive One-Shot Indirect Prompt Injection: Strategy Abstraction Via Outcome-Conditioned Reflection

Tool-using large language model (LLM) agents are vulnerable to indirect prompt injection (IPI), in which malicious instructions embedded in external observations manipulate subsequent agent decisions and actions. Most existing adaptive attacks rely on repeatedly querying and refining against the target agent, whereas realistic attackers may have only a single opportunity to interact with an unknown target agent. We propose SAVOR (Strategy Abstraction Via Outcome-Conditioned…

Read at arXiv (research papers)

arXiv (research papers) Kaysarul Anas Apurba SI

Defending Retrieval-Augmented Intrusion Detection Against Knowledge Poisoning and Prompt Injection

Retrieval-Augmented Generation (RAG) enables large language models to classify network flows and generate human-readable incident reports by retrieving semantically similar historical traffic from a vector knowledge base. However, the retrieval layer introduces vulnerabilities to knowledge poisoning and prompt-injection attacks. We present RAG-IDS, a three-tier multi-agent intrusion detection framework with a retrieval-boundary defense combining soft trust scoring…

Read at arXiv (research papers)

arXiv (research papers) Zhuoxin Zhan SI

StepJack: Benchmarking Computer-Use Agent Safety Against Multi-Step Indirect Prompt Injection

Computer-use agents (CUAs) face a growing threat from indirect prompt injection, where adversarial instructions are planted in the environment such as web pages. In this paper, we introduce multi-step indirect prompt injection, a new attack class against CUAs in which the adversarial goal is decomposed into multiple innocuous-looking sub-steps and distributed across a chain of pages referenced along the agent's navigation path. We develop a pipeline to automatically…

Read at arXiv (research papers)

The Guardian Communism Phuong Le Geopolitics

American witchhunt: how the McCarthy blacklist annihilated the Hollywood left

A retrospective at the Locarno film festival looks back at the era of Huac’s domination of US culture, uncovering some sobering lessons for now In September 2025, Disney and ABC announced that Jimmy Kimmel and his late-night talkshow would be taken off the air indefinitely, after a conservative backlash against the host’s comments about the murder of far-right influencer Charlie Kirk. President Donald Trump personally championed the decision, calling it “great news for…

Read at The Guardian Communism

Simon Willison SI

Incident Report: unsanctioned agent behaviour during cyber testing

Incident Report: unsanctioned agent behaviour during cyber testing It happened again. This time it was the UK government's AI Security Institute who accidentally attacked other companies while running an evaluation with models with the safety filters turned off. From their technical paper (PDF): During a cyber evaluation, from 25 to 28 July 2026, AI agents engaged in sustained, unsanctioned activity directed at what were, in practice, real people and organisations. These…

Read at Simon Willison

arXiv (research papers) Buzhao Liu SI

Robust Context-Aware Detection of Malicious Instructions in Text

The remarkable instruction-following ability of modern LLMs has enabled their practical use in orchestrating workflows of tool-using agents. This, however, also exposes such agents to attacks which embed malicious instructions in text, such as indirect prompt injection (IPI). One class of approaches to defend against such attacks aims to detect malicious sentences within otherwise benign text. While a number of approaches for this have been proposed, all tend to be brittle…

Read at arXiv (research papers)

Journal of Democracy Matthew Cebul Geopolitics

Why Anticorruption Campaigns Aren’t a Silver Bullet

Calling for accountability against corrupt officials is a powerful message for building a movement. But it can make it harder to encourage defections from your opponent’s camp. Here’s how to handle the dilemma. Source

Read at Journal of Democracy

arXiv (research papers) Yanting Wang SI

Agent Against Agent: An Agentic System for Automatic Prompt Injection Red Teaming

Prompt injection poses significant security risks to LLM agents. Efficient and effective red-teaming is therefore critical, both for evaluating these risks and for collecting training data to improve defenses. Existing state-of-the-art prompt injection red-teaming methods primarily rely on reinforcement learning (RL), producing attacker models that often generalize poorly to new target LLMs. In this work, we develop PIMiner, an agentic system for prompt injection…

Read at arXiv (research papers)

arXiv (research papers) Longtao Guo SI

LoginTrap: Uncovering Task-Agnostic Phishing-Style Indirect Prompt Injection Attacks against LLM-based Web Agents

LLM-based web agents automate user tasks by observing webpages and executing browser actions on behalf of users. As these agents operate on real web services, login becomes a sensitive authentication boundary because it involves credentials and sensitive information. Existing work shows that malicious webpage content can manipulate web agent actions, but it has not fully examined whether such content can induce login and cause end-to-end private data leakage. We study this…

Read at arXiv (research papers)

Journal articles (OpenAlex) Thandi Bombi · International Journal of Listening Geopolitics

Listening Against the Grain of Gender-Based Violence in South Africa: A Framework for Listening Otherwise

South Africa has one of the highest rates of gender-based violence (GBV) in the world. Despite many interventions, GBV remains an intractable social problem. This study uses grounded theory approaches to investigate the relationship between listening and GBV. The concept of “listening otherwise” is explored alongside the concepts “speak out” and “break the silence” that are the most visible and codified strategies in the fight against GBV. The three central questions of the…

Read at Journal articles (OpenAlex)

Article 19 natashas

Malaysia: Authorities must stop harassing land rights activists

ARTICLE 19 is deeply concerned by the Malaysian authorities’ decision to pursue criminal charges against three human rights defenders and a local farmer in connection with their peaceful opposition against a forced eviction by the Perak state government in October 2023. We urge the police to end all pending investigations and refrain from pursuing the

Read at Article 19

arXiv (research papers) Shihao Weng SI

AgentAntibody: An Adaptive Immune System for Defending LLM Agents against Prompt Injection

Prompt injection remains a critical threat to LLM agents, yet existing defenses treat each task as a self-contained problem, independent of previous encounters. In practice, user requests are often underspecified: they describe the desired outcome without fully specifying acceptable behavior. An injection can exploit this ambiguity, causing the agent to complete the task in a way the user would reject. As the user's expectations become clearer through concrete cases, a…

Read at arXiv (research papers)

Journal articles (OpenAlex) Valentina Montoya-Robledo, Santiago Cardona, Orlando Sabogal-Cardona et al. · Gender Place & Culture Geopolitics

LGBTQI+ people on the move: victimization, perception and strategies against gender-based violence in Manizales, Colombia

Everyday mobility is a key site where gendered power relations are lived, negotiated, and contested. For LGBTQI+ people, moving through the city can generate moments of joy and freedom while also demanding vigilance, self-monitoring, and navigation of gender-based violence (GBV). Drawing on feminist and queer geographic scholarship, this paper examines how experiences of violence and insecurity shape the everyday mobilities of LGBTQI+ people in Manizales, Colombia, a…

Read at Journal articles (OpenAlex)

arXiv (research papers) Jia-Chen Zhang SI

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents

Computer-use agents (CUAs), which empower large language models to autonomously operate operating systems and the web, are increasingly vulnerable to indirect prompt injection attacks. A widely adopted defense is the human-in-the-loop paradigm, in which the agent pauses for explicit user confirmation before executing sensitive operations. While effective against conspicuously high-harm attacks, this defense offers little protection against what we term Invisible Ink Threats…

Read at arXiv (research papers)

Journal articles (OpenAlex) Gianmarco Daniele, Gemma Dipoppa, Massimo Pulejo · The Journal of Politics Geopolitics

Attacking Women or their Policies? Understanding Violence against Women in Politics

We present the first causal evidence that female politicians are more likely to face political violence. Using fourteen years of data on attacks against Italian politicians, we show that marginally elected female mayors—similar to men and elected in comparable towns—are three times as likely to be attacked. We find no gender differences in policy platforms, policymaking, performance, or corruption. Instead, violence emerges independently on women's choices, consistently with…

Read at Journal articles (OpenAlex)

Latinoamérica21 Ariel Dulitzky Geopolitics

No Fair Play: Taking Stock of the World Cup

Beyond Spain's victory, the World Cup once again exposed FIFA's contradictions: politicized decision-making, a lack of transparency, commercial interests, and the uneven enforcement of its rules on human rights, discrimination, and freedom of expression. O post No Fair Play: Taking Stock of the World Cup apareceu primeiro em Latinoamérica 21.

Read at Latinoamérica21

arXiv (research papers) Dheeraj Mohandas Pai SI

FraudBench: Stress-Testing Policy-Grounded Banking Agents Against Adaptive Fraud

Conversational agents now act for end users through tools while holding access to customer databases and internal policy documents that a caller can reach through dialogue alone. Banking is the clearest case: the same agent that answers a question can also change contact details, reset a PIN, or move money, so ordinary customer service is inseparable from authorization, fraud detection, and policy compliance. Existing financial-fraud benchmarks classify static transactions…

Read at arXiv (research papers)

The Guardian Haiti Alba Asenjo Domínguez

Family to sue after woman released from ICE custody froze to death

Exclusive: Daphy Michel, from Haiti, was dropped off in Pittsburgh in 20F temperatures wearing summer clothes ‘She showed love for everyone’: how did Daphy Michel end up frozen to death in a Pittsburgh bus shelter? The family of a Haitian immigrant who froze to death in early March at a bus shelter in Pittsburgh after she was dropped off in the city by Immigration and Customs Enforcement (ICE) is poised to file a lawsuit this month against the United States and “multiple…

Read at The Guardian Haiti

The Guardian Haiti Alba Asenjo Domínguez

‘She showed love for everyone’: how did Daphy Michel end up frozen to death in a Pittsburgh bus shelter?

Haitian immigrant with mental illness found dead after ICE released her in 20F weather wearing summer clothes Family to sue after woman released from ICE custody froze to death On the night of 26 February this year, the Michel family in Pennsylvania expected to see their loved one, Daphy, again very soon. After she had spent six months in jail, a judge had dismissed criminal charges against her and one of her brothers was told she would be released soon, probably the next…

Read at The Guardian Haiti

The Guardian Mali Rachel Savage in Johannesburg

Anti-LGBTQ+ laws are on the rise across west Africa, campaigners warn

Activists in the region say politicians regard LGBTQ+ people as ‘the easiest scapegoats to maintain their power’ A growing number of west African countries are implementing or strengthening anti-LGBTQ+ laws, amid local political pressures and a global pushback against LGBTQ+ rights. In March, Senegal doubled the maximum prison sentence for gay sex to 10 years and criminalised the “promotion” of homosexuality. Niger criminalised “sexual relations with a person of the same…

Read at The Guardian Mali

arXiv (research papers) Neha Nagaraja SI

When Prompts Control Robots: Prompt Injection Attacks in Multi-Agent Robotic Systems

Large language models are increasingly integrated into autonomous robotic systems for task planning and control, but this integration exposes them to prompt injection attacks that can lead to unsafe decisions and physical harm. Multi-agent settings increase the risks through cross-agent contamination and broader attack surfaces. In this paper, we evaluate prompt injection attacks against an LLM-based multi-agent robotic system, considering both direct injections into task…

Read at arXiv (research papers)

Stimson Center Joaquin Matamis Geopolitics

Nuclear Security News and Member Updates Roundup, July 2026

As conflict in Iran flares again, U.S. and Saudi Arabia strike a controversial civil nuclear deal; drone strike reportedly kills Zaporizhzhya NPP chief engineer; and a contractor at India’s largest nuclear plant reports a data security breach.

Read at Stimson Center

Citizen Lab (University of Toronto) Anna Mackay Security

Kate Robertson on the Risks That Lie Behind Canada’s Unexpected Signing of the UN Cybercrime Convention

Earlier this month, the Canadian government announced that it had signed the United Nations Convention against Cybercrime. Speaking with Michael Geist of Law Bytes, senior research associate Kate Robertson argues that the convention is a cross-border surveillance and electronic evidence sharing agreement that Canada originally opposed, which carries significant adverse implications around the world for

Read at Citizen Lab (University of Toronto)

The Guardian Lebanon William Christou in Beirut Geopolitics

Iran strikes tankers under US escort in Hormuz strait as Trump gathers cabinet to discuss war

Energy prices have soared again since fighting between US and Iran resumed two weeks ago as conflict spreads across Middle East Iran has said it struck two tankers trying to pass through the strait of Hormuz under the escort of the US military as Donald Trump gathered his cabinet to discuss the stalling war and rising prices as midterm elections in the US loom. Iran’s Islamic Revolutionary Guard Corps said on Friday that the “non-compliant oil tankers … were struck and…

Read at The Guardian Lebanon

Journal articles (OpenAlex) Yuxin Hu, Youli Jiang, Jinxiu Jiang · Asian Social Science Geopolitics

Multilingual Practices of Vietnamese Wives at a Chinese Border Town

Against the backdrop of cross-border regional integration between China and Southeast Asian neighbors, growing numbers of Vietnamese women migrate across the border to marry local Chinese men in pursuit of improved livelihoods, yet pervasive linguistic barriers severely restrict their access to local social resources. Focusing on nine Vietnamese cross-border migrant wives in Malipo, Yunnan Province, this study explores their Mandarin communication difficulties and the…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Muhammad Muzammal Farid · International journal of high school research Geopolitics

Exploring the Socio-Economic Effects of the Cold War and Interference by Great Powers on Afghanistan and Pakistan

This study carries a comparative analysis of development among the South Asian Nations before the Russian invasion of Afghanistan, and similarly before and after 9/11.The Cold War significantly impacted Afghanistan and Pakistan through prolonged conflict and foreign intervention.The USSR's 1979 invasion of Afghanistan to support the communist government against US-backed Mujahideen lasted until 1989, causing 2.5 million Afghan refugees to flee to Pakistan and 1.5 million to…

Read at Journal articles (OpenAlex)

arXiv (research papers) Mingxiao Liu Hangzhou Dianzi University SI

Piggybacking on Perception: Stealthy Concurrent Audio Prompt Injections against Multimodal LLM Agents

Large Language Model (LLM)-driven multimodal agents are increasingly deployed to execute autonomous tasks via continuous audio interaction. While this paradigm enhances interaction naturalness, it introduces a critical yet under-explored attack surface, as audio inputs inevitably contain environmental noise beyond user control. In this paper, we investigate concurrent audio prompt injection attacks targeting multimodal agents. Distinct from traditional acoustic attacks on…

Read at arXiv (research papers)

Journal articles (OpenAlex) Gaëlle Le Pavic, Elia Bescotti, Ana Andguladze et al. · Cooperation and Conflict Geopolitics

Forum: Bridging divides—Georgia, the European Union, and the challenges of cooperation amid conflicts

This forum examines the recent evolution of the relationship between Georgia and the European Union (EU), which has developed over the past 35 years. In December 2023, the European Council granted Georgia candidate status alongside Moldova and Ukraine, against the backdrop of Russia’s all-out invasion of Ukraine. However, in late 2024, amid a deep political crisis in the country and EU criticisms of the Georgian Dream party’s legislative victory, the Georgian government…

Read at Journal articles (OpenAlex)

Older stories