Research Papers

This paper investigates the complex interplay between AI developers, regulators, users, and the media in fostering trustworthy AI systems. Using evolutionary game theory and large language models (LLMs), we model the strategic interactions among these actors under different regulatory regimes. The research explores two key mechanisms for achieving responsible governance, safe AI development and adoption of safe AI: incentivising effective regulation through media reporting, and conditioning user trust on commentariats' recommendation. The findings highlight the crucial role of the media in providing information to users, potentially acting as a form of "soft" regulation by investigating developers or regulators, as a substitute to institutional AI regulation (which is still absent in many regions). Both game-theoretic analysis and LLM-based simulations reveal conditions under which effective regulation and trustworthy AI development emerge, emphasising the importance of considering the influence of different regulatory regimes from an evolutionary game-theoretic perspective. The study concludes that effective governance requires managing incentives and costs for high quality commentaries.

AI expansion has accelerated workplace adoption of new technologies. Yet, it is unclear whether and how knowledge workers are supported and trained to safely use AI. Inadequate training may lead to unrealized benefits if workers abandon tools, or perpetuate biases if workers misinterpret AI-based outcomes. In a workshop with 39 workers from 26 countries specializing in human resources, labor law, standards creation, and worker training, we explored questions and ideas they had about safely adopting AI. We held 17 follow-up interviews to further investigate what skills and training knowledge workers need to achieve safe and effective AI in practice. We synthesize nine training topics participants surfaced for knowledge workers related to challenges around understanding what AI is, misinterpreting outcomes, exacerbating biases, and worker rights. We reflect how these training topics might be addressed under different contexts, imagine HCI research prototypes as potential training tools, and consider ways to ensure training does not perpetuate harmful values.

Explainable AI for Clinical Outcome Prediction: A Survey of Clinician Perceptions and Preferences

Jun Hou, Lucy Lu Wang

Explainable AI (XAI) techniques are necessary to help clinicians make sense of AI predictions and integrate predictions into their decision-making workflow. In this work, we conduct a survey study to understand clinician preference among different XAI techniques when they are used to interpret model predictions over text-based EHR data. We implement four XAI techniques (LIME, Attention-based span highlights, exemplar patient retrieval, and free-text rationales generated by LLMs) on an outcome prediction model that uses ICU admission notes to predict a patient's likelihood of experiencing in-hospital mortality. Using these XAI implementations, we design and conduct a survey study of 32 practicing clinicians, collecting their feedback and preferences on the four techniques. We synthesize our findings into a set of recommendations describing when each of the XAI techniques may be more appropriate, their potential limitations, as well as recommendations for improvement.

Explainable Artificial Intelligence (XAI) techniques hold significant potential for enhancing the causal discovery process, which is crucial for understanding complex systems in areas like healthcare, economics, and artificial intelligence. However, no causal discovery methods currently incorporate explainability into their models to derive the causal graphs. Thus, in this paper we explore this innovative approach, as it offers substantial potential and represents a promising new direction worth investigating. Specifically, we introduce ReX, a causal discovery method that leverages machine learning (ML) models coupled with explainability techniques, specifically Shapley values, to identify and interpret significant causal relationships among variables. Comparative evaluations on synthetic datasets comprising continuous tabular data reveal that ReX outperforms state-of-the-art causal discovery methods across diverse data generation processes, including non-linear and additive noise models. Moreover, ReX was tested on the Sachs single-cell protein-signaling dataset, achieving a precision of 0.952 and recovering key causal relationships with no incorrect edges. Taking together, these results showcase ReX's effectiveness in accurately recovering true causal structures while minimizing false positive predictions, its robustness across diverse datasets, and its applicability to real-world problems. By combining ML and explainability techniques with causal discovery, ReX bridges the gap between predictive modeling and causal inference, offering an effective tool for understanding complex causal structures.

New research focuses on creating artificial intelligence (AI) solutions for network intrusion detection systems (NIDS), drawing its inspiration from the ever-growing number of intrusions on networked systems, increasing its complexity and intelligibility. Hence, the use of explainable AI (XAI) techniques in real-world intrusion detection systems comes from the requirement to comprehend and elucidate black-box AI models to security analysts. In an effort to meet such requirements, this paper focuses on applying and evaluating White-Box XAI techniques (particularly LRP, IG, and DeepLift) for NIDS via an end-to-end framework for neural network models, using three widely used network intrusion datasets (NSL-KDD, CICIDS-2017, and RoEduNet-SIMARGL2021), assessing its global and local scopes, and examining six distinct assessment measures (descriptive accuracy, sparsity, stability, robustness, efficiency, and completeness). We also compare the performance of white-box XAI methods with black-box XAI methods. The results show that using White-box XAI techniques scores high in robustness and completeness, which are crucial metrics for IDS. Moreover, the source codes for the programs developed for our XAI evaluation framework are available to be improved and used by the research community.

Background: Despite the current ubiquity of Large Language Models (LLMs) across the medical domain, there is a surprising lack of studies which address their reasoning behaviour. We emphasise the importance of understanding reasoning behaviour as opposed to high-level prediction accuracies, since it is equivalent to explainable AI (XAI) in this context. In particular, achieving XAI in medical LLMs used in the clinical domain will have a significant impact across the healthcare sector. Results: Therefore, in this work, we adapt the existing concept of reasoning behaviour and articulate its interpretation within the specific context of medical LLMs. We survey and categorise current state-of-the-art approaches for modeling and evaluating reasoning reasoning in medical LLMs. Additionally, we propose theoretical frameworks which can empower medical professionals or machine learning engineers to gain insight into the low-level reasoning operations of these previously obscure models. We also outline key open challenges facing the development of Large Reasoning Models. Conclusion: The subsequent increased transparency and trust in medical machine learning models by clinicians as well as patients will accelerate the integration, application as well as further development of medical AI for the healthcare system as a whole.

As AI systems become increasingly powerful, the need for safe AI has become more pressing. Humans are an attractive model for AI safety: as the only known agents capable of general intelligence, they perform robustly even under conditions that deviate significantly from prior experiences, explore the world safely, understand pragmatics, and can cooperate to meet their intrinsic goals. Intelligence, when coupled with cooperation and safety mechanisms, can drive sustained progress and well-being. These properties are a function of the architecture of the brain and the learning algorithms it implements. Neuroscience may thus hold important keys to technical AI safety that are currently underexplored and underutilized. In this roadmap, we highlight and critically evaluate several paths toward AI safety inspired by neuroscience: emulating the brain's representations, information processing, and architecture; building robust sensory and motor systems from imitating brain data and bodies; fine-tuning AI systems on brain data; advancing interpretability using neuroscience methods; and scaling up cognitively-inspired architectures. We make several concrete recommendations for how neuroscience can positively impact AI safety.

The remarkable capabilities of Large Language Models (LLMs) make them increasingly compelling for adoption in real-world healthcare applications. However, the risks associated with using LLMs in medical applications have not been systematically characterized. We propose using five key principles for safe and trustworthy medical AI: Truthfulness, Resilience, Fairness, Robustness, and Privacy, along with ten specific aspects. Under this comprehensive framework, we introduce a novel MedGuard benchmark with 1,000 expert-verified questions. Our evaluation of 11 commonly used LLMs shows that the current language models, regardless of their safety alignment mechanisms, generally perform poorly on most of our benchmarks, particularly when compared to the high performance of human physicians. Despite recent reports indicate that advanced LLMs like ChatGPT can match or even exceed human performance in various medical tasks, this study underscores a significant safety gap, highlighting the crucial need for human oversight and the implementation of AI safety guardrails.

SafetyAnalyst: Interpretable, Transparent, and Steerable Safety Moderation for AI Behavior

Jing-Jing Li, Valentina Pyatkin, Max Kleiman-Weiner +7

The ideal AI safety moderation system would be both structurally interpretable (so its decisions can be reliably explained) and steerable (to align to safety standards and reflect a community's values), which current systems fall short on. To address this gap, we present SafetyAnalyst, a novel AI safety moderation framework. Given an AI behavior, SafetyAnalyst uses chain-of-thought reasoning to analyze its potential consequences by creating a structured "harm-benefit tree," which enumerates harmful and beneficial actions and effects the AI behavior may lead to, along with likelihood, severity, and immediacy labels that describe potential impacts on stakeholders. SafetyAnalyst then aggregates all effects into a harmfulness score using 28 fully interpretable weight parameters, which can be aligned to particular safety preferences. We applied this framework to develop an open-source LLM prompt safety classification system, distilled from 18.5 million harm-benefit features generated by frontier LLMs on 19k prompts. On comprehensive benchmarks, we show that SafetyAnalyst (average F1=0.81) outperforms existing moderation systems (average F1$<$0.72) on prompt safety classification, while offering the additional advantages of interpretability, transparency, and steerability.

This study is located in the Human-Centered Artificial Intelligence (HCAI) and focuses on the results of a user-centered assessment of commonly used eXplainable Artificial Intelligence (XAI) algorithms, specifically investigating how humans understand and interact with the explanations provided by these algorithms. To achieve this, we employed a multi-disciplinary approach that included state-of-the-art research methods from social sciences to measure the comprehensibility of explanations generated by a state-of-the-art lachine learning model, specifically the Gradient Boosting Classifier (XGBClassifier). We conducted an extensive empirical user study involving interviews with 39 participants from three different groups, each with varying expertise in data science, data visualization, and domain-specific knowledge related to the dataset used for training the machine learning model. Participants were asked a series of questions to assess their understanding of the model's explanations. To ensure replicability, we built the model using a publicly available dataset from the UC Irvine Machine Learning Repository, focusing on edible and non-edible mushrooms. Our findings reveal limitations in existing XAI methods and confirm the need for new design principles and evaluation techniques that address the specific information needs and user perspectives of different classes of AI stakeholders. We believe that the results of our research and the cross-disciplinary methodology we developed can be successfully adapted to various data types and user profiles, thus promoting dialogue and address opportunities in HCAI research. To support this, we are making the data resulting from our study publicly available.

Item Response Theory (IRT) has been widely used in educational psychometrics to assess student ability, as well as the difficulty and discrimination of test questions. In this context, discrimination specifically refers to how effectively a question distinguishes between students of different ability levels, and it does not carry any connotation related to fairness. In recent years, IRT has been successfully used to evaluate the predictive performance of Machine Learning (ML) models, but this paper marks its first application in fairness evaluation. In this paper, we propose a novel Fair-IRT framework to evaluate a set of predictive models on a set of individuals, while simultaneously eliciting specific parameters, namely, the ability to make fair predictions (a feature of predictive models), as well as the discrimination and difficulty of individuals that affect the prediction results. Furthermore, we conduct a series of experiments to comprehensively understand the implications of these parameters for fairness evaluation. Detailed explanations for item characteristic curves (ICCs) are provided for particular individuals. We propose the flatness of ICCs to disentangle the unfairness between individuals and predictive models. The experiments demonstrate the effectiveness of this framework as a fairness evaluation tool. Two real-world case studies illustrate its potential application in evaluating fairness in both classification and regression tasks. Our paper aligns well with the Responsible Web track by proposing a Fair-IRT framework to evaluate fairness in ML models, which directly contributes to the development of a more inclusive, equitable, and trustworthy AI.

The growing capabilities of AI models are leading to their wider use, including in safety-critical domains. Explainable AI (XAI) aims to make these models safer to use by making their inference process more transparent. However, current explainability methods are seldom evaluated in the way they are intended to be used: by real-world end users. To address this, we conducted a large-scale user study with 85 healthcare practitioners in the context of human-AI collaborative chest X-ray analysis. We evaluated three types of explanations: visual explanations (saliency maps), natural language explanations, and a combination of both modalities. We specifically examined how different explanation types influence users depending on whether the AI advice and explanations are factually correct. We find that text-based explanations lead to significant over-reliance, which is alleviated by combining them with saliency maps. We also observe that the quality of explanations, that is, how much factually correct information they entail, and how much this aligns with AI correctness, significantly impacts the usefulness of the different explanation types.