Large language models (LLMs) are being increasingly integrated into everyday products and services, such as coding tools and writing assistants. As these embedded AI applications are deployed globally, there is a growing concern that the AI models underlying these applications prioritize Western values. This paper investigates what happens when a Western-centric AI model provides writing suggestions to users from a different cultural background. We conducted a cross-cultural controlled experiment with 118 participants from India and the United States who completed culturally grounded writing tasks with and without AI suggestions. Our analysis reveals that AI provided greater efficiency gains for Americans compared to Indians. Moreover, AI suggestions led Indian participants to adopt Western writing styles, altering not just what is written but also how it is written. These findings show that Western-centric AI models homogenize writing toward Western norms, diminishing nuances that differentiate cultural expression.

Large multilingual models have significantly advanced natural language processing (NLP) research. However, their high resource demands and potential biases from diverse data sources have raised concerns about their effectiveness across low-resource languages. In contrast, monolingual models, trained on a single language, may better capture the nuances of the target language, potentially providing more accurate results. This study benchmarks the cross-lingual transfer capabilities from a high-resource language to a low-resource language for both, monolingual and multilingual models, focusing on Kinyarwanda and Kirundi, two Bantu languages. We evaluate the performance of transformer based architectures like Multilingual BERT (mBERT), AfriBERT, and BantuBERTa against neural-based architectures such as BiGRU, CNN, and char-CNN. The models were trained on Kinyarwanda and tested on Kirundi, with fine-tuning applied to assess the extent of performance improvement and catastrophic forgetting. AfriBERT achieved the highest cross-lingual accuracy of 88.3% after fine-tuning, while BiGRU emerged as the best-performing neural model with 83.3% accuracy. We also analyze the degree of forgetting in the original language post-fine-tuning. While monolingual models remain competitive, this study highlights that multilingual models offer strong cross-lingual transfer capabilities in resource limited settings.

Stereotypes are generalised assumptions about societal groups, and even state-of-the-art LLMs using in-context learning struggle to identify them accurately. Due to the subjective nature of stereotypes, where what constitutes a stereotype can vary widely depending on cultural, social, and individual perspectives, robust explainability is crucial. Explainable models ensure that these nuanced judgments can be understood and validated by human users, promoting trust and accountability. We address these challenges by introducing HEARTS (Holistic Framework for Explainable, Sustainable, and Robust Text Stereotype Detection), a framework that enhances model performance, minimises carbon footprint, and provides transparent, interpretable explanations. We establish the Expanded Multi-Grain Stereotype Dataset (EMGSD), comprising 57,201 labeled texts across six groups, including under-represented demographics like LGBTQ+ and regional stereotypes. Ablation studies confirm that BERT models fine-tuned on EMGSD outperform those trained on individual components. We then analyse a fine-tuned, carbon-efficient ALBERT-V2 model using SHAP to generate token-level importance values, ensuring alignment with human understanding, and calculate explainability confidence scores by comparing SHAP and LIME outputs. Finally, HEARTS is applied to assess stereotypical bias in 12 LLM outputs, revealing a gradual reduction in bias over time within model families.

Earth System Models (ESMs) are essential for understanding the interaction between human activities and the Earth's climate. However, the computational demands of ESMs often limit the number of simulations that can be run, hindering the robust analysis of risks associated with extreme weather events. While low-cost climate emulators have emerged as an alternative to emulate ESMs and enable rapid analysis of future climate, many of these emulators only provide output on at most a monthly frequency. This temporal resolution is insufficient for analyzing events that require daily characterization, such as heat waves or heavy precipitation. We propose using diffusion models, a class of generative deep learning models, to effectively downscale ESM output from a monthly to a daily frequency. Trained on a handful of ESM realizations, reflecting a wide range of radiative forcings, our DiffESM model takes monthly mean precipitation or temperature as input, and is capable of producing daily values with statistical characteristics close to ESM output. Combined with a low-cost emulator providing monthly means, this approach requires only a small fraction of the computational resources needed to run a large ensemble. We evaluate model behavior using a number of extreme metrics, showing that DiffESM closely matches the spatio-temporal behavior of the ESM output it emulates in terms of the frequency and spatial characteristics of phenomena such as heat waves, dry spells, or rainfall intensity.

As Artificial Intelligence applications expand, the evaluation of models faces heightened scrutiny. Ensuring public readiness requires evaluation datasets, which differ from training data by being disjoint and ethically sourced in compliance with privacy regulations. The performance and fairness of face recognition systems depend significantly on the quality and representativeness of these evaluation datasets. This data is sometimes scraped from the internet without user's consent, causing ethical concerns that can prohibit its use without proper releases. In rare cases, data is collected in a controlled environment with consent, however, this process is time-consuming, expensive, and logistically difficult to execute. This creates a barrier for those unable to conjure the immense resources required to gather ethically sourced evaluation datasets. To address these challenges, we introduce the Synthetic Identity Generation pipeline, or SIG, that allows for the targeted creation of ethical, balanced datasets for face recognition evaluation. Our proposed and demonstrated pipeline generates high-quality images of synthetic identities with controllable pose, facial features, and demographic attributes, such as race, gender, and age. We also release an open-source evaluation dataset named ControlFace10k, consisting of 10,008 face images of 3,336 unique synthetic identities balanced across race, gender, and age, generated using the proposed SIG pipeline. We analyze ControlFace10k along with a non-synthetic BUPT dataset using state-of-the-art face recognition algorithms to demonstrate its effectiveness as an evaluation tool. This analysis highlights the dataset's characteristics and its utility in assessing algorithmic bias across different demographic groups.

Most of the ancient underground aqueducts that enabled humans to settle in the world's hottest and driest regions have been lost over time. Now, archaeologists are rediscovering them by using artificial intelligence to analyse spy satellite images taken during the cold war. The oldest known underground aqueducts that are found across much of North Africa and the Middle East are called qanats and are up to 3000 years old.

In the past couple of years, regulators have been caught off guard again and again as tech companies compete to launch ever more advanced AI models. As it stands, it seems there's little anyone can do to delay or prevent the release of a model that poses excessive risks. Existing measures to mitigate AI risks aren't enough to protect us, so we need new approaches. One could be a kind of AI safety hotline tasked with expert volunteers. Read more about how the hotline could work.

Hosted by Eleanor Drage and Kerry Mackereth, The Good Robot is a podcast which explores the many complex intersections between gender, feminism and technology. In this episode, we talk to Daniel Leufer and Caterina Rodelli from Access Now, a global advocacy organization that focuses on the impact of the digital on human rights. As leaders in this field, they've been working hard to ensure that the European Union's AI Act doesn't undermine human rights or indeed fundamental democratic values. They share with us how the EU AI act was put together, the Act's particular downfalls, and where the opportunities are for us as citizens or as digital rights activists to get involved and make sure that it's upheld by companies across the world. Note: this episode was recorded back in February 2024.

Retrieval-Augmented Generation (RAG) has quickly grown into a pivotal paradigm in the development of Large Language Models (LLMs). While much of the current research in this field focuses on performance optimization, particularly in terms of accuracy and efficiency, the trustworthiness of RAG systems remains an area still under exploration. From a positive perspective, RAG systems are promising to enhance LLMs by providing them with useful and up-to-date knowledge from vast external databases, thereby mitigating the long-standing problem of hallucination. While from a negative perspective, RAG systems are at the risk of generating undesirable contents if the retrieved information is either inappropriate or poorly utilized. To address these concerns, we propose a unified framework that assesses the trustworthiness of RAG systems across six key dimensions: factuality, robustness, fairness, transparency, accountability, and privacy. Within this framework, we thoroughly review the existing literature on each dimension. Additionally, we create the evaluation benchmark regarding the six dimensions and conduct comprehensive evaluations for a variety of proprietary and open-source models. Finally, we identify the potential challenges for future research based on our investigation results. Through this work, we aim to lay a structured foundation for future investigations and provide practical insights for enhancing the trustworthiness of RAG systems in real-world applications.

Hunger crises are critical global issues affecting millions, particularly in low-income and developing countries. This research investigates how machine learning can be utilized to predict and inform decisions regarding famine and hunger crises. By leveraging a diverse set of variables (natural, economic, and conflict-related), three machine learning models (Linear Regression, XGBoost, and RandomForestRegressor) were employed to predict food consumption scores, a key indicator of household nutrition. The RandomForestRegressor emerged as the most accurate model, with an average prediction error of 10.6%, though accuracy varied significantly across countries, ranging from 2% to over 30%. Notably, economic indicators were consistently the most significant predictors of average household nutrition, while no single feature dominated across all regions, underscoring the necessity for comprehensive data collection and tailored, country-specific models. These findings highlight the potential of machine learning, particularly Random Forests, to enhance famine prediction, suggesting that continued research and improved data gathering are essential for more effective global hunger forecasting.