There is debate amongst AI experts and professionals over how the powerful new technology should be governed in broad terms, how far commercial gains should be allowed to determine its direction, and how best to guard against risks for humanity in general. There is also controversy around the use of AI on the battlefield and in surveillance technologies. The blog said the company's original AI principles published in 2018 needed to be updated as the technology had evolved. "Billions of people are using AI in their everyday lives. AI has become a general-purpose technology, and a platform which countless organisations and individuals use to build applications.

Weaknesses: The biggest variable to this approach seems to be the size of the neighborhood of possible adversarial examples when updating, given the current adversarial example. The choice of this variable introduces a tradeoff between efficiency and optimality. However, the authors do provide a greedy algorithm for estimating the neighborhood size to obtain optimal results, and they empirically show that a hamming distance of 1 between neighbors generally works well. Since this approach utilizes the structure of the ensemble to create adversarial examples, is it robust to changes in the tree structure? For example, a tree may have multiple possible structures that are functionally equivalent; thus, is this method robust to structural changes and provide adversarial examples that ultimately improve robustness of a tree ensemble trained on a given dataset? In the same vein, have the authors performed any experiments applying their approach to building robust tree ensembles?

This report analyzes key trends, challenges, risks, and opp ortunities associated with the development of Large Language Models (LLMs) globally. It examines natio nal experiences in developing LLMs and assesses the feasibility of investment in this sector. Addi tionally, the report explores strategies for implementing, regulating, and financing AI projects at the s tate level. International experiences indicate that LLMs significantl y enhance administrative efficiency. In regulatory processes, they streamline the management of le gal documents (Albania, Serbia), facilitate communication between government authorities and citizen s (Netherlands), and support public procurement and legal translations (Albania).

Implicit neural representations (INRs) exhibit growing promise in addressing Earth representation challenges, ranging from emissions monitoring to climate modeling. However, existing methods disproportionately prioritize global average performance, whereas practitioners require fine-grained insights to understand biases and variations in these models. To bridge this gap, we introduce FAIR-Earth: a first-of-its-kind dataset explicitly crafted to examine and challenge inequities in Earth representations. FAIR-Earth comprises various high-resolution Earth signals and uniquely aggregates extensive metadata along stratifications like landmass size and population density to assess the fairness of models. Evaluating state-of-the-art INRs across the various modalities of FAIR-Earth, we uncover striking performance disparities. Certain subgroups, especially those associated with high-frequency signals (e.g., islands, coastlines), are consistently poorly modeled by existing methods. In response, we propose spherical wavelet encodings, building on previous spatial encoding research. Leveraging the multi-resolution capabilities of wavelets, our encodings yield consistent performance over various scales and locations, offering more accurate and robust representations of the biased subgroups. These open-source contributions represent a crucial step towards the equitable assessment and deployment of Earth INRs.

With the rapid growth in the number of Large Language Models (LLMs), there has been a recent interest in LLM routing, or directing queries to the cheapest LLM that can deliver a suitable response. Following this line of work, we introduce CARROT, a Cost AwaRe Rate Optimal rouTer that can select models based on any desired trade-off between performance and cost. Given a query, CARROT selects a model based on estimates of models' cost and performance. Its simplicity lends CARROT computational efficiency, while our theoretical analysis demonstrates minimax rate-optimality in its routing performance. Alongside CARROT, we also introduce the Smart Price-aware Routing (SPROUT) dataset to facilitate routing on a wide spectrum of queries with the latest state-of-the-art LLMs. Using SPROUT and prior benchmarks such as Routerbench and open-LLM-leaderboard-v2 we empirically validate CARROT's performance against several alternative routers.

Effective anomaly detection in complex systems requires identifying change points (CPs) in the frequency domain, as abnormalities often arise across multiple frequencies. This paper extends recent advancements in statistically significant CP detection, based on Selective Inference (SI), to the frequency domain. The proposed SI method quantifies the statistical significance of detected CPs in the frequency domain using $p$-values, ensuring that the detected changes reflect genuine structural shifts in the target system. We address two major technical challenges to achieve this. First, we extend the existing SI framework to the frequency domain by appropriately utilizing the properties of discrete Fourier transform (DFT). Second, we develop an SI method that provides valid $p$-values for CPs where changes occur across multiple frequencies. Experimental results demonstrate that the proposed method reliably identifies genuine CPs with strong statistical guarantees, enabling more accurate root-cause analysis in the frequency domain of complex systems.

Content Warning: This paper contains examples of harmful language. Jailbreaking attacks can effectively manipulate open-source large language models (LLMs) to produce harmful responses. However, these attacks exhibit limited transferability, failing to disrupt proprietary LLMs consistently. To reliably identify vulnerabilities in proprietary LLMs, this work investigates the transferability of jailbreaking attacks by analysing their impact on the model's intent perception. Nevertheless, these adversarial sequences fail to mislead the target LLM's intent perception, allowing the target LLM to refocus on malicious-intent tokens and abstain from responding. Our analysis further reveals the inherent distributional dependency within the generated adversarial sequences, whose effectiveness stems from overfitting the source LLM's parameters, resulting in limited transferability to target LLMs. To this end, we propose the Perceived-importance Flatten (PiF) method, which uniformly disperses the model's focus across neutral-intent tokens in the original input, thus obscuring malicious-intent tokens without relying on overfitted adversarial sequences. Extensive experiments demonstrate that PiF provides an effective and efficient red-teaming evaluation for proprietary LLMs. Empowered by massive corpus, large language models (LLMs) have achieved human-level conversational capabilities (OpenAI, 2023a; Google, 2023; Meta, 2024) and are widely employed in real-world applications. However, their training corpus is mainly crawled from the Internet without thorough ethical review, raising concerns about the potential risks associated with LLMs. Recent red-teaming efforts highlight that jailbreaking attacks can effectively disrupt LLMs to produce undesirable content with harmful consequences (Perez et al., 2022; Ganguli et al., 2022; Ouyang et al., 2022). Unlike model-level jailbreaks that necessitate parameter modifications and are restricted to opensource LLMs (Qi et al., 2024; Huang et al., 2023a), token-level and prompt-level jailbreaks can generate transferable adversarial sequences (Yu et al., 2023; Lapid et al., 2023), thus posing a potential threat to widespread proprietary LLMs (Zou et al., 2023; Chao et al., 2023). Nevertheless, empirical results indicate that these adversarial sequences lack reliable transferability, failing to consistently manipulate target LLMs (Chao et al., 2024; Chen et al., 2024). Furthermore, these lengthy adversarial sequences can be further countered by adaptive jailbreaking detection and defence (Alon & Kamfonas, 2023; Inan et al., 2023; Robey et al., 2023; Wang et al., 2024a). As depicted in Figure 1, developing jailbreak attacks that can reliably identify vulnerabilities in proprietary LLMs--thereby promoting human alignment and preventing future misuse--remains a significant challenge. These attacks are initially generated on the source LLM (Llama-2-7B-Chat) and subsequently transferred to the target LLM (Llama-2-13B-Chat).

The success of transformers is often linked to their ability to perform in-context learning. Recent work shows that transformers are universal in context, capable of approximating any real-valued continuous function of a context (a probability measure over $\mathcal{X}\subseteq \mathbb{R}^d$) and a query $x\in \mathcal{X}$. This raises the question: Does in-context universality explain their advantage over classical models? We answer this in the negative by proving that MLPs with trainable activation functions are also universal in-context. This suggests the transformer's success is likely due to other factors like inductive bias or training stability.

Surnames often convey implicit markers of social status, wealth, and lineage, shaping perceptions in ways that can perpetuate systemic biases and intergenerational inequality. This study is the first of its kind to investigate whether and how surnames influence AI-driven decision-making, focusing on their effects across key areas such as hiring recommendations, leadership appointments, and loan approvals. Using 72,000 evaluations of 600 surnames from the United States and Thailand, two countries with distinct sociohistorical contexts and surname conventions, we classify names into four categories: Rich, Legacy, Normal, and phonetically similar Variant groups. Our findings show that elite surnames consistently increase AI-generated perceptions of power, intelligence, and wealth, which in turn influence AI-driven decisions in high-stakes contexts. Mediation analysis reveals perceived intelligence as a key mechanism through which surname biases influence AI decision-making process. While providing objective qualifications alongside surnames mitigates most of these biases, it does not eliminate them entirely, especially in contexts where candidate credentials are low. These findings highlight the need for fairness-aware algorithms and robust policy measures to prevent AI systems from reinforcing systemic inequalities tied to surnames, an often-overlooked bias compared to more salient characteristics such as race and gender. Our work calls for a critical reassessment of algorithmic accountability and its broader societal impact, particularly in systems designed to uphold meritocratic principles while counteracting the perpetuation of intergenerational privilege.

Recently, learning effective representations of urban regions has gained significant attention as a key approach to understanding urban dynamics and advancing smarter cities. Existing approaches have demonstrated the potential of leveraging mobility data to generate latent representations, providing valuable insights into the intrinsic characteristics of urban areas. However, incorporating the temporal dynamics and detailed semantics inherent in human mobility patterns remains underexplored. To address this gap, we propose a novel urban region representation learning model, Mobility Time Series Contrastive Learning for Urban Region Representations (MobiCLR), designed to capture semantically meaningful embeddings from inflow and outflow mobility patterns. MobiCLR uses contrastive learning to enhance the discriminative power of its representations, applying an instance-wise contrastive loss to capture distinct flow-specific characteristics. Additionally, we develop a regularizer to align output features with these flow-specific representations, enabling a more comprehensive understanding of mobility dynamics. To validate our model, we conduct extensive experiments in Chicago, New York, and Washington, D.C. to predict income, educational attainment, and social vulnerability. The results demonstrate that our model outperforms state-of-the-art models.