If you were asked to visualise a stereotypical British man, what would you think of? According to AI, the answer is an overweight man wearing a football shirt. Instagram account @reimagineuk asked AI to create videos of the most stereotypical men around the world - with hilarious results. While the British man looks casual in his football shirt, men from other countries are depicted with fancier outfits. The stereotypical man from Portugal sports a white shirt and a waistcoat, while the man from Nigeria can be seen wearing a bright orange suit.

While machine learning has significantly advanced Network Intrusion Detection Systems (NIDS), particularly within IoT environments where devices generate large volumes of data and are increasingly susceptible to cyber threats, these models remain vulnerable to adversarial attacks. Our research reveals a critical flaw in existing adversarial attack methodologies: the frequent violation of domain-specific constraints, such as numerical and categorical limits, inherent to IoT and network traffic. This leads to up to 80.3% of adversarial examples being invalid, significantly overstating real-world vulnerabilities. These invalid examples, though effective in fooling models, do not represent feasible attacks within practical IoT deployments. Consequently, relying on these results can mislead resource allocation for defense, inflating the perceived susceptibility of IoT-enabled NIDS models to adversarial manipulation. Furthermore, we demonstrate that simpler surrogate models like Multi-Layer Perceptron (MLP) generate more valid adversarial examples compared to complex architectures such as CNNs and LSTMs. Using the MLP as a surrogate, we analyze the transferability of adversarial severity to other ML/DL models commonly used in IoT contexts. This work underscores the importance of considering both domain constraints and model architecture when evaluating and designing robust ML/DL models for security-critical IoT and network applications.

This paper describes the mwBTFreddy dataset, a resource developed to support flash flood damage assessment in urban Malawi, specifically focusing on the impacts of Cyclone Freddy in 2023. The dataset comprises paired pre- and post-disaster satellite images sourced from Google Earth Pro, accompanied by JSON files containing labelled building annotations with geographic coordinates and damage levels (no damage, minor, major, or destroyed). Developed by the Kuyesera AI Lab at the Malawi University of Business and Applied Sciences, this dataset is intended to facilitate the development of machine learning models tailored to building detection and damage classification in African urban contexts. It also supports flood damage visualisation and spatial analysis to inform decisions on relocation, infrastructure planning, and emergency response in climate-vulnerable regions.

Zero-Day Botnet Attack Detection in IoV: A Modular Approach Using Isolation Forests and Particle Swarm Optimization Abdelaziz Amara korba 2, Nour Elislem Karabadji 1, and Y acine Ghamri-Doudane 2 1 National Higher School of T echnology and Engineering, LTSE, E3360100, Annaba, Algeria. 2 L3I, University of La Rochelle, France Abstract --The Internet of V ehicles (IoV) is transforming transportation by enhancing connectivity and enabling autonomous driving. However, this increased interconnectivity introduces new security vulnerabilities. Bot malware and cyberattacks pose significant risks to Connected and Autonomous V ehicles (CA Vs), as demonstrated by real-world incidents involving remote vehicle system compromise. T o address these challenges, we propose an edge-based Intrusion Detection System (IDS) that monitors network traffic to and from CA Vs. Our detection model is based on a meta-ensemble classifier capable of recognizing known (N-day) attacks and detecting previously unseen (zero-day) attacks. The approach involves training multiple Isolation Forest (IF) models on Multi-access Edge Computing (MEC) servers, with each IF specialized in identifying a specific type of botnet attack. These IFs, either trained locally or shared by other MEC nodes, are then aggregated using a Particle Swarm Optimization (PSO) based stacking strategy to construct a robust meta-classifier . The proposed IDS has been evaluated on a vehicular botnet dataset, achieving an average detection rate of 92.80% for N-day attacks and 77.32% for zero-day attacks.

Right now, most big AI labs have a team figuring out ways that rogue AIs might escape supervision, or secretly collude with each other against humans. But there's a more mundane way we could lose control of civilisation: we might simply become obsolete. This wouldn't require any hidden plots – if AI and robotics keep improving, it's what happens by default. Well, AI developers are firmly on track to build better replacements for humans in almost every role we play: not just economically as workers and decision-makers, but culturally as artists and creators, and even socially as friends and romantic companions. What place will humans have when AI can do everything we do, only better?

Sudan's army says the paramilitary Rapid Support Forces (RSF) attacked a military airbase and other facilities in the vicinity of Port Sudan airport. The army said on Sunday that the airbase was targeted using a drone, as well as a cargo warehouse and some civilian facilities, in the first attack in the eastern city by the RSF. There are reports of some damage after drones hit an ammunition depot. "Both the civilian and military airports are in the same place. What we know from residents in the port city is that five drones were launched by the RSF and targeted the airbase," Al Jazeera's Hiba Morgan said, reporting from the capital, Khartoum.

The proliferation of generative AI and deceptive synthetic media threatens the global information ecosystem, especially across the Global Majority. This report from WITNESS highlights the limitations of current AI detection tools, which often underperform in real-world scenarios due to challenges related to explainability, fairness, accessibility, and contextual relevance. In response, WITNESS introduces the Truly Innovative and Effective AI Detection (TRIED) Benchmark, a new framework for evaluating detection tools based on their real-world impact and capacity for innovation. Drawing on frontline experiences, deceptive AI cases, and global consultations, the report outlines how detection tools must evolve to become truly innovative and relevant by meeting diverse linguistic, cultural, and technological contexts. It offers practical guidance for developers, policy actors, and standards bodies to design accountable, transparent, and user-centered detection solutions, and incorporate sociotechnical considerations into future AI standards, procedures and evaluation frameworks. By adopting the TRIED Benchmark, stakeholders can drive innovation, safeguard public trust, strengthen AI literacy, and contribute to a more resilient global information credibility.

Bora Chung's Red Sword is set on a disputed planet While there are no big names publishing new science fiction novels this May, there are some real gems nonetheless – including a big tip from me, Grace Chan's near-future Every Version of You. I want to press it into the hands of everyone I know. There are also two fascinating sci-fi-edged thrillers out this month, by Adam Oyebanji and Barnaby Martin, while Catherine Chidgey's creepy The Book of Guilt has intrigued me enough to make it my next read – if it's not ousted by Bora Chung's real history-inspired story of war on an alien planet, Red Sword, that is… Set in late-21st-century Australia, this novel (published in Australia in 2022 but out now more widely) follows Tao-Yi in a world where most people spend their lives in an immersive virtual reality called Gaia. Every morning, she climbs into a pod in her apartment to enter Gaia, where she works and socialises. In the real world, the unrelenting heat of the sun means there are no trees left and hardly any animals: this is a terrifying vision of the future.

Racial disparities in healthcare expenditures are well-documented, yet the underlying drivers remain complex and require further investigation. This study employs causal and counterfactual path-specific effects to quantify how various factors, including socioeconomic status, insurance access, health behaviors, and health status, mediate these disparities. Using data from the Medical Expenditures Panel Survey, we estimate how expenditures would differ under counterfactual scenarios in which the values of specific mediators were aligned across racial groups along selected causal pathways. A key challenge in this analysis is ensuring robustness against model misspecification while addressing the zero-inflation and right-skewness of healthcare expenditures. For reliable inference, we derive asymptotically linear estimators by integrating influence function-based techniques with flexible machine learning methods, including super learners and a two-part model tailored to the zero-inflated, right-skewed nature of healthcare expenditures.

--Privacy in federated learning is crucial, encompassing two key aspects: safeguarding the privacy of clients' data and maintaining the privacy of the federator's objective from the clients. While the first aspect has been extensively studied, the second has received much less attention. We present a novel approach that addresses both concerns simultaneously, drawing inspiration from techniques in knowledge distillation and private information retrieval to provide strong information-theoretic privacy guarantees. Traditional private function computation methods could be used here; however, they are typically limited to linear or polynomial functions. T o overcome these constraints, our approach unfolds in three stages. In stage 0, clients perform the necessary computations locally. In stage 1, these results are shared among the clients, and in stage 2, the federator retrieves its desired objective without compromising the privacy of the clients' data. The crux of the method is a carefully designed protocol that combines secret-sharing-based multi-party computation and a graph-based private information retrieval scheme. We show that our method outperforms existing tools from the literature when properly adapted to this setting. We consider federated learning (FL), a framework where a federator and a set of clients with private data collaborate to train a neural network. Due to privacy constraints, the clients' data cannot be directly shared with the federator or among the clients. This privacy concern has been extensively studied in the literature [2]-[6]. There exists a second, often overlooked, privacy concern: ensuring the privacy of the federator's objective used to train the neural network. This aspect has not been explored in the literature to the same extent. We present a novel approach that ensures the privacy of the clients' data and simultaneously hides the objective of the federator through a careful combination of a secure aggregation method and a tailored private information retrieval (PIR) scheme. This project is funded by DFG (German Research Foundation) projects under Grant Agreement Nos. Part of the work was done when RB and ME visited RU at EPFL supported in parts by EuroTech Visiting Researcher Programme grants.