Russia has launched a large-scale drone and missile attack on Ukraine's capital Kyiv, killing several people and wounding more than a hundred others according to local officials. Al Jazeera's Assed Baig has been to one of the attack sites there.

Mark Zuckerberg announced in April that the company would make huge capital expenditures in the coming year to keep up in the race to develop cutting-edge artificial intelligence. He made good on that promise last week with a 15bn "AI superintelligence" team that would feature reported nine-figure salaries and a 49% investment in Scale AI. Before Meta's investment, Scale counted most of the major players in AI among its clients, and some of them were less than thrilled with the development. Bloomberg puts it succinctly: Scale AI's Wang Brings to Meta Knowledge of What Everyone Else is Doing. Google, Scale's largest customer, got scared.

Former Foreign Service officer advocates for social media vetting of visa applicants, explaining how AI could enhance screening to identify anti-American sentiments.

For example, a few weeks before my trip, the Trump administration rescinded Biden's executive order on AI safety and DOGE began turning to AI to decide which federal programs to cut. Then, more recently, House Republicans passed a 10-year moratorium on US states' ability to regulate AI (though it has yet to be passed by the Senate). What all this points to is a new reality in the United States where responsible AI is no longer a priority (if it ever genuinely was). But this has also made me think more deeply about the stakes of deploying AI in situations that directly affect human lives, and about what success would even look like. When Amsterdam's welfare department began developing the algorithm that became Smart Check, the municipality followed virtually every recommendation in the responsible-AI playbook: consulting external experts, running bias tests, implementing technical safeguards, and seeking stakeholder feedback.

Waves of Russian missile and drone strikes have killed at least 15 people and injured 116 others, with most of the casualties in Kyiv, Ukrainian officials have reported. The massive aerial assault overnight into Tuesday struck 27 locations in the Ukrainian capital, damaging residential buildings and critical infrastructure, according to Interior Minister Ihor Klymenko. Ukrainian officials were quick to call for international attention on the attacks as Kyiv pushes diplomatic efforts to raise pressure on Moscow to agree a ceasefire. "Today, the enemy spared neither drones nor missiles," Klymenko said, describing the attack as one of the largest against Kyiv since Russia launched its full-scale invasion of the country in February 2022. Thirty apartments were destroyed in a single residential block, and emergency services were searching through the rubble for possible survivors, Klymenko added.

Artificial intelligence could spark an accidental nuclear war, conflict experts fear. The Stockholm International Peace Research Institute (SIPRI), the world's leading organisation on nuclear assessments, said technologies like AI are aggravating the risk carried with growing global nuclear stockpiles. SIPRI pointed to China's rapidly growing stockpile, from 500 to 600 in a single year, as well as the imminent expiry of the final arms control treaty between the US and Russia, two nuclear-armed nations. The institute's director, Dan Smith, warned: 'One component of the coming arms race will be the attempt to gain and maintain a competitive edge in artificial intelligence (AI), both for offensive and defensive purposes. 'There are benefits to be found but the careless adoption of AI could significantly increase nuclear risk.'

We study the memory complexity of machine unlearning algorithms that provide strong data deletion guarantees to the users. Formally, consider an algorithm for a particular learning task that initially receives a training dataset. Then, after learning, it receives data deletion requests from a subset of users (of arbitrary size), and the goal of unlearning is to perform the task as if the learner never received the data of deleted users. In this paper, we ask how many bits of storage are needed to be able to delete certain training samples at a later time. We focus on the task of realizability testing, where the goal is to check whether the remaining training samples are realizable within a given hypothesis class \(\mathcal{H}\). Toward that end, we first provide a negative result showing that the VC dimension is not a characterization of the space complexity of unlearning. In particular, we provide a hypothesis class with constant VC dimension (and Littlestone dimension), but for which any unlearning algorithm for realizability testing needs to store \(Ω(n)\)-bits, where \(n\) denotes the size of the initial training dataset. In fact, we provide a stronger separation by showing that for any hypothesis class \(\mathcal{H}\), the amount of information that the learner needs to store, so as to perform unlearning later, is lower bounded by the \textit{eluder dimension} of \(\mathcal{H}\), a combinatorial notion always larger than the VC dimension. We complement the lower bound with an upper bound in terms of the star number of the underlying hypothesis class, albeit in a stronger ticketed-memory model proposed by Ghazi et al. (2023). Since the star number for a hypothesis class is never larger than its Eluder dimension, our work highlights a fundamental separation between central and ticketed memory models for machine unlearning.

With the rapid advancement of aerospace technology and the large-scale deployment of low Earth orbit (LEO) satellite constellations, the challenges facing astronomical observations and deep space exploration have become increasingly pronounced. As a result, the demand for high-precision orbital data on space objects-along with comprehensive analyses of satellite positioning, constellation configurations, and deep space satellite dynamics-has grown more urgent. However, there remains a notable lack of publicly accessible, real-world datasets to support research in areas such as space object maneuver behavior prediction and collision risk assessment. This study seeks to address this gap by collecting and curating a representative dataset of maneuvering behavior from Starlink satellites. The dataset integrates Two-Line Element (TLE) catalog data with corresponding high-precision ephemeris data, thereby enabling a more realistic and multidimensional modeling of space object behavior. It provides valuable insights into practical deployment of maneuver detection methods and the evaluation of collision risks in increasingly congested orbital environments.

The increasing prevalence of large language models (LLMs) is influencing global value systems. However, these models frequently exhibit a pronounced WEIRD (Western, Educated, Industrialized, Rich, Democratic) cultural bias due to lack of attention to minority values. This monocultural perspective may reinforce dominant values and marginalize diverse cultural viewpoints, posing challenges for the development of equitable and inclusive AI systems. In this work, we introduce a systematic framework designed to boost fair and robust cross-cultural consensus among LLMs. We model consensus as a Nash Equilibrium and employ a game-theoretic negotiation method based on Policy-Space Response Oracles (PSRO) to simulate an organized cross-cultural negotiation process. To evaluate this approach, we construct regional cultural agents using data transformed from the World Values Survey (WVS). Beyond the conventional model-level evaluation method, We further propose two quantitative metrics, Perplexity-based Acceptence and Values Self-Consistency, to assess consensus outcomes. Experimental results indicate that our approach generates consensus of higher quality while ensuring more balanced compromise compared to baselines. Overall, it mitigates WEIRD bias by guiding agents toward convergence through fair and gradual negotiation steps.

Machine unlearning focuses on efficiently removing specific data from trained models, addressing privacy and compliance concerns with reasonable costs. Although exact unlearning ensures complete data removal equivalent to retraining, it is impractical for large-scale models, leading to growing interest in inexact unlearning methods. However, the lack of formal guarantees in these methods necessitates the need for robust evaluation frameworks to assess their privacy and effectiveness. In this work, we first identify several key pitfalls of the existing unlearning evaluation frameworks, e.g., focusing on average-case evaluation or targeting random samples for evaluation, incomplete comparisons with the retraining baseline. Then, we propose RULI (Rectified Unlearning Evaluation Framework via Likelihood Inference), a novel framework to address critical gaps in the evaluation of inexact unlearning methods. RULI introduces a dual-objective attack to measure both unlearning efficacy and privacy risks at a per-sample granularity. Our findings reveal significant vulnerabilities in state-of-the-art unlearning methods, where RULI achieves higher attack success rates, exposing privacy risks underestimated by existing methods. Built on a game-based foundation and validated through empirical evaluations on both image and text data (spanning tasks from classification to generation), RULI provides a rigorous, scalable, and fine-grained methodology for evaluating unlearning techniques.