Artificial Intelligence is rapidly embedding itself within militaries, economies, and societies, reshaping their very foundations. Given the depth and breadth of its consequences, it has never been more pressing to understand how to ensure that AI systems are safe, ethical, and have a positive societal impact. This book aims to provide a comprehensive approach to understanding AI risk. Our primary goals include consolidating fragmented knowledge on AI risk, increasing the precision of core ideas, and reducing barriers to entry by making content simpler and more comprehensible. The book has been designed to be accessible to readers from diverse backgrounds. You do not need to have studied AI, philosophy, or other such topics. The content is skimmable and somewhat modular, so that you can choose which chapters to read. We introduce mathematical formulas in a few places to specify claims more precisely, but readers should be able to understand the main points without these.

A yet unmet challenge in algorithmic fairness is the problem of intersectionality, that is, achieving fairness across the intersection of multiple groups -- and verifying that such fairness has been attained. Because intersectional groups tend to be small, verifying whether a model is fair raises statistical as well as moral-methodological challenges. This paper (1) elucidates the problem of intersectionality in algorithmic fairness, (2) develops desiderata to clarify the challenges underlying the problem and guide the search for potential solutions, (3) illustrates the desiderata and potential solutions by sketching a proposal using simple hypothesis testing, and (4) evaluates, partly empirically, this proposal against the proposed desiderata.

Most Deontic Logics take obligation as primitive and leave the others as derived from obligations. On the other hand, normative reasoning/legal theory identifies two different notions of permission: Strong Permission and Weak Permission. While the definitions of the types of permission vary, and other notions of permission have been proposed (for a discussion on the topic, see Hansson (2013)), often strong permission is taken as a derogation to a prohibition or the obligation to the contrary, and we have a weak permission when we fail to obtain the obligation of the contrary. Another way to look at the issue is whether there are norms that explicitly permit something. If there are and the norms are effective, then we obtain an explicit (strong) permission.

In recent years, text-to-image (T2I) generation models have made significant progress in generating high-quality images that align with text descriptions. However, these models also face the risk of unsafe generation, potentially producing harmful content that violates usage policies, such as explicit material. Existing safe generation methods typically focus on suppressing inappropriate content by erasing undesired concepts from visual representations, while neglecting to sanitize the textual representation. Although these methods help mitigate the risk of misuse to certain extent, their robustness remains insufficient when dealing with adversarial attacks. Given that semantic consistency between input text and output image is a fundamental requirement for T2I models, we identify that textual representations (i.e., prompt embeddings) are likely the primary source of unsafe generation. To this end, we propose a vision-agnostic safe generation framework, Embedding Sanitizer (ES), which focuses on erasing inappropriate concepts from prompt embeddings and uses the sanitized embeddings to guide the model for safe generation. ES is applied to the output of the text encoder as a plug-and-play module, enabling seamless integration with different T2I models as well as other safeguards. In addition, ES's unique scoring mechanism assigns a score to each token in the prompt to indicate its potential harmfulness, and dynamically adjusts the sanitization intensity to balance defensive performance and generation quality. Through extensive evaluation on five prompt benchmarks, our approach achieves state-of-the-art robustness by sanitizing the source (prompt embedding) of unsafe generation compared to nine baseline methods. It significantly outperforms existing safeguards in terms of interpretability and controllability while maintaining generation quality.

Heatwaves (HWs) are extreme atmospheric events that produce significant societal and environmental impacts. Predicting these extreme events remains challenging, as their complex interactions with large-scale atmospheric and climatic variables are difficult to capture with traditional statistical and dynamical models. This work presents a general method for driver identification in extreme climate events. A novel framework (STCO-FS) is proposed to identify key immediate (short-term) HW drivers by combining clustering algorithms with an ensemble evolutionary algorithm. The framework analyzes spatio-temporal data, reduces dimensionality by grouping similar geographical nodes for each variable, and develops driver selection in spatial and temporal domains, identifying the best time lags between predictive variables and HW occurrences. The proposed method has been applied to analyze HWs in the Adda river basin in Italy. The approach effectively identifies significant variables influencing HWs in this region. This research can potentially enhance our understanding of HW drivers and predictability.

The demands for accurate and representative generative AI systems means there is an increased demand on participatory evaluation structures. While these participatory structures are paramount to to ensure non-dominant values, knowledge and material culture are also reflected in AI models and the media they generate, we argue that dominant structures of community participation in AI development and evaluation are not explicit enough about the benefits and harms that members of socially marginalized groups may experience as a result of their participation. Without explicit interrogation of these benefits by AI developers, as a community we may remain blind to the immensity of systemic change that is needed as well. To support this provocation, we present a speculative case study, developed from our own collective experiences as AI researchers. We use this speculative context to itemize the barriers that need to be overcome in order for the proposed benefits to marginalized communities to be realized, and harms mitigated.

As data-driven and AI-based decision making gains widespread adoption in most disciplines, it is crucial that both data privacy and decision fairness are appropriately addressed. While differential privacy (DP) provides a robust framework for guaranteeing privacy and several widely accepted methods have been proposed for improving fairness, the vast majority of existing literature treats the two concerns independently. For methods that do consider privacy and fairness simultaneously, they often only apply to a specific machine learning task, limiting their generalizability. In response, we introduce SAFES, a Sequential PrivAcy and Fairness Enhancing data Synthesis procedure that sequentially combines DP data synthesis with a fairness-aware data transformation. SAFES allows full control over the privacy-fairness-utility trade-off via tunable privacy and fairness parameters. We illustrate SAFES by combining AIM, a graphical model-based DP data synthesizer, with a popular fairness-aware data pre-processing transformation. Empirical evaluations on the Adult and COMPAS datasets demonstrate that for reasonable privacy loss, SAFES-generated synthetic data achieve significantly improved fairness metrics with relatively low utility loss.

The past few years have witnessed an unprecedented improvement in the capabilities of Large Language Models (LLMs), driven by the success in scaling up autoregressive language modeling in terms of data, model size, and the amount of compute used for training (Kaplan et al., 2020). LLMs have demonstrated exceptional linguistic abilities (Brown, 2020; Achiam et al., 2023), general tool use (Schick et al., 2024; Cai et al., 2023), and commonsense reasoning (Wei et al., 2022; OpenAI, 2024), among other impressive capabilities. The success of LLMs as general-purpose assistants motivates research and development to extend instruction-tuning to the vision-language multimodal space (Liu et al., 2023; Gemini Team, 2023). These vision-language multimodal models, which can process and generate both text and images, also achieve human-expert performance on a wide range of tasks, such as (document) visual question answering (Antol et al., 2015; Mathew et al., 2021), image captioning (Lin et al., 2014), and image-text retrieval (Plummer et al., 2015). While these vision-language multimodal models hold tremendous promise for many applications, they should be used along with proper system guardrails to ensure safe and responsible deployment, because they can generate or propagate harmful content when interacting with online users. However, most existing guardrails (Inan et al., 2023; Llama Team, 2024b,a; Yuan et al., 2024; Ghosh et al., 2024) for the interaction (e.g., conversation) between humans and AI agents are text-only: conversation data involving other modalities, such as images, cannot be used as inputs for such guardrails. This calls for a safeguard tool for classifying safety risks in prompts and responses for conversations with multimodal contents involved. In this work, we introduce Llama Guard 3 Vision, a multimodal LLM-based safeguard for human-AI conversations that involves image understanding: it can be used to safeguard content for both mutimodal LLM inputs (prompt classification) and mutimodal LLM responses (response classification). Unlike text-only Llama Guard versions (Inan et al., 2023; Llama Team, 2024b,a), it is specifically designed to support image reasoning use cases and is optimized to detect harmful multimodal (text and image) prompts and text responses to these prompts.

There is an urgent need to identify both short and long-term risks from newly emerging types of Artificial Intelligence (AI), as well as available risk management measures. In response, and to support global efforts in regulating AI and writing safety standards, we compile an extensive catalog of risk sources and risk management measures for general-purpose AI (GPAI) systems, complete with descriptions and supporting examples where relevant. This work involves identifying technical, operational, and societal risks across model development, training, and deployment stages, as well as surveying established and experimental methods for managing these risks. To the best of our knowledge, this paper is the first of its kind to provide extensive documentation of both GPAI risk sources and risk management measures that are descriptive, self-contained and neutral with respect to any existing regulatory framework. This work intends to help AI providers, standards experts, researchers, policymakers, and regulators in identifying and mitigating systemic risks from GPAI systems. For this reason, the catalog is released under a public domain license for ease of direct use by stakeholders in AI governance and standards.

The promise of human behavioral simulation--general-purpose computational agents that replicate human behavior across domains--could enable broad applications in policymaking and social science. We present a novel agent architecture that simulates the attitudes and behaviors of 1,052 real individuals--applying large language models to qualitative interviews about their lives, then measuring how well these agents replicate the attitudes and behaviors of the individuals that they represent. The generative agents replicate participants' responses on the General Social Survey 85% as accurately as participants replicate their own answers two weeks later, and perform comparably in predicting personality traits and outcomes in experimental replications. Our architecture reduces accuracy biases across racial and ideological groups compared to agents given demographic descriptions. This work provides a foundation for new tools that can help investigate individual and collective behavior.