Despite their high predictive accuracies, current machine learning systems often exhibit systematic biases stemming from annotation artifacts or insufficient support for certain classes in the dataset. Recent work proposes automatic methods for identifying and explaining systematic biases using keywords. We introduce DISCERN, a framework for interpreting systematic biases in text classifiers using language explanations. DISCERN iteratively generates precise natural language descriptions of systematic errors by employing an interactive loop between two large language models. Finally, we use the descriptions to improve classifiers by augmenting classifier training sets with synthetically generated instances or annotated examples via active learning. On three text-classification datasets, we demonstrate that language explanations from our framework induce consistent performance improvements that go beyond what is achievable with exemplars of systematic bias. Finally, in human evaluations, we show that users can interpret systematic biases more effectively (by over 25% relative) and efficiently when described through language explanations as opposed to cluster exemplars.

Accurate economic simulations often require many experimental runs, particularly when combined with reinforcement learning. Unfortunately, training reinforcement learning agents in multi-agent economic environments can be slow. This paper introduces EconoJax, a fast simulated economy, based on the AI economist. EconoJax, and its training pipeline, are completely written in JAX. This allows EconoJax to scale to large population sizes and perform large experiments, while keeping training times within minutes. Through experiments with populations of 100 agents, we show how real-world economic behavior emerges through training within 15 minutes, in contrast to previous work that required several days. To aid and inspire researchers to build more rich and dynamic economic simulations, we open-source EconoJax on Github at: https://github.com/ponseko/econojax.

Artificial Intelligence (AI) is a field that utilizes computing and often, data and statistics, intensively together to solve problems or make predictions. AI has been evolving with literally unbelievable speed over the past few years, and this has led to an increase in social, cultural, industrial, scientific, and governmental concerns about the ethical development and use of AI systems worldwide. The ASA has issued a statement on ethical statistical practice and AI (ASA, 2024), which echoes similar statements from other groups. Here we discuss the support for ethical statistical practice and ethical AI that has been established in long-standing human rights law and ethical practice standards for computing and statistics. There are multiple sources of support for ethical statistical practice and ethical AI deriving from these source documents, which are critical for strengthening the operationalization of the "Statement on Ethical AI for Statistics Practitioners". These resources are explicated for interested readers to utilize to guide their development and use of AI in, and through, their statistical practice.

Artificial intelligence (AI) technology has been evolving more rapidly over the past decade. With new ML models, data sources, and increased computational power, AI researchers have developed AI technologies that can understand language, recognize and create images and videos, program, and make scientific inferences. Recent advances in advanced AI technologies have evolved beyond traditional narrow domain AI to approximate or exceed artificial general intelligence (AGI) based on large language models (LLMs) or foundation models (FMs). These advanced AI systems are performing at or above human levels in complex problem solving, sophisticated natural language processing, and multi-domain tasks, and have the potential to revolutionize a wide range of fields, including science, industry, healthcare, and education. They are already surpassing human capabilities in certain task domains, such as Go, strategy games, and protein folding prediction [1] [2]. For these reasons, concerns about the safety and trustworthiness of advanced AI are growing rapidly alongside its development. The increasing complexity and autonomy of advanced AI systems is raising concerns that they could lead to new forms of safety and security risks, such as (1) uncontrollability, (2) conflicts with human values in ethical decision-making, (3) long-term socioeconomic impacts, and (4) safety assurance. In response, international standardization efforts are underway to ensure the safety and trustworthiness of advanced AI. By developing internationally agreed technical standards, efforts are being made to apply consistent safety and trustworthiness criteria to the development and use of advanced AI systems and minimize potential risks.

The European Union's Artificial Intelligence Act (AI Act) introduces comprehensive guidelines for the development and oversight of Artificial Intelligence (AI) and Machine Learning (ML) systems, with significant implications for Graph Neural Networks (GNNs). This paper addresses the unique challenges posed by the AI Act for GNNs, which operate on complex graph-structured data. The legislation's requirements for data management, data governance, robustness, human oversight, and privacy necessitate tailored strategies for GNNs. Our study explores the impact of these requirements on GNN training and proposes methods to ensure compliance. We provide an in-depth analysis of bias, robustness, explainability, and privacy in the context of GNNs, highlighting the need for fair sampling strategies and effective interpretability techniques. Our contributions fill the research gap by offering specific guidance for GNNs under the new legislative framework and identifying open questions and future research directions.

Generative models such as Large Language Models (LLM) and Multimodal Large Language models (MLLMs) trained on massive web corpora can memorize and disclose individuals' confidential and private data, raising legal and ethical concerns. While many previous works have addressed this issue in LLM via machine unlearning, it remains largely unexplored for MLLMs. To tackle this challenge, we introduce Multimodal Large Language Model Unlearning Benchmark (MLLMU-Bench), a novel benchmark aimed at advancing the understanding of multimodal machine unlearning. MLLMU-Bench consists of 500 fictitious profiles and 153 profiles for public celebrities, each profile feature over 14 customized question-answer pairs, evaluated from both multimodal (image+text) and unimodal (text) perspectives. The benchmark is divided into four sets to assess unlearning algorithms in terms of efficacy, generalizability, and model utility. Finally, we provide baseline results using existing generative model unlearning algorithms. Surprisingly, our experiments show that unimodal unlearning algorithms excel in generation and cloze tasks, while multimodal unlearning approaches perform better in classification tasks with multimodal inputs.

Ensuring the safety of large language model (LLM) applications is essential for developing trustworthy artificial intelligence. Current LLM safety benchmarks have two limitations. First, they focus solely on either discriminative or generative evaluation paradigms while ignoring their interconnection. Second, they rely on standardized inputs, overlooking the effects of widespread prompting techniques, such as system prompts, few-shot demonstrations, and chain-of-thought prompting. To overcome these issues, we developed SG-Bench, a novel benchmark to assess the generalization of LLM safety across various tasks and prompt types. This benchmark integrates both generative and discriminative evaluation tasks and includes extended data to examine the impact of prompt engineering and jailbreak on LLM safety. Our assessment of 3 advanced proprietary LLMs and 10 open-source LLMs with the benchmark reveals that most LLMs perform worse on discriminative tasks than generative ones, and are highly susceptible to prompts, indicating poor generalization in safety alignment. We also explain these findings quantitatively and qualitatively to provide insights for future research.

Authorship attribution aims to identify the origin or author of a document. Traditional approaches have heavily relied on manual features and fail to capture long-range correlations, limiting their effectiveness. Recent advancements leverage text embeddings from pre-trained language models, which require significant fine-tuning on labeled data, posing challenges in data dependency and limited interpretability. Large Language Models (LLMs), with their deep reasoning capabilities and ability to maintain long-range textual associations, offer a promising alternative. This study explores the potential of pre-trained LLMs in one-shot authorship attribution, specifically utilizing Bayesian approaches and probability outputs of LLMs. Our methodology calculates the probability that a text entails previous writings of an author, reflecting a more nuanced understanding of authorship. By utilizing only pre-trained models such as Llama-3-70B, our results on the IMDb and blog datasets show an impressive 85\% accuracy in one-shot authorship classification across ten authors. Our findings set new baselines for one-shot authorship analysis using LLMs and expand the application scope of these models in forensic linguistics. This work also includes extensive ablation studies to validate our approach.

Linguistically inclusive LLMs -- which maintain good performance regardless of the language with which they are prompted -- are necessary for the diffusion of AI benefits around the world. Multilingual jailbreaks that rely on language translation to evade safety measures undermine the safe and inclusive deployment of AI systems. We provide policy recommendations to enhance the multilingual capabilities of AI while mitigating the risks of multilingual jailbreaks. We examine how a language's level of resourcing relates to how vulnerable LLMs are to multilingual jailbreaks in that language. We do this by testing five advanced AI models across 24 official languages of the EU. Building on prior research, we propose policy actions that align with the EU legal landscape and institutional framework to address multilingual jailbreaks, while promoting linguistic inclusivity. These include mandatory assessments of multilingual capabilities and vulnerabilities, public opinion research, and state support for multilingual AI development. The measures aim to improve AI safety and functionality through EU policy initiatives, guiding the implementation of the EU AI Act and informing regulatory efforts of the European AI Office.

Machine learning models are often used to automate or support decisions in applications such as lending and hiring. In such settings, consumer protection rules mandate that we provide a list of "principal reasons" to consumers who receive adverse decisions. In practice, lenders and employers identify principal reasons by returning the top-scoring features from a feature attribution method. In this work, we study how such practices align with one of the underlying goals of consumer protection - recourse - i.e., educating individuals on how they can attain a desired outcome. We show that standard attribution methods can mislead individuals by highlighting reasons without recourse - i.e., by presenting consumers with features that cannot be changed to achieve recourse. We propose to address these issues by scoring features on the basis of responsiveness - i.e., the probability that an individual can attain a desired outcome by changing a specific feature. We develop efficient methods to compute responsiveness scores for any model and any dataset under complex actionability constraints. We present an extensive empirical study on the responsiveness of explanations in lending and demonstrate how responsiveness scores can be used to construct feature-highlighting explanations that lead to recourse and mitigate harm by flagging instances with fixed predictions.