Artificial Intelligence for Natural Hazards Risk Analysis: Potential,Challenges, and Research Needs

Artificial intelligence (AI) methods have seen increasingly widespread use in everything from consumer products and driverless cars to fraud detection and weather forecasting. The use of AI has transformed many of these application domains. There are ongoing efforts at leveraging AI for disaster risk analysis. This article takes a critical look at the use of AI for disaster risk analysis. What is the potential? How is the use of AI in this field different from its use in nondisaster fields? What challenges need to be overcome for this potential to be realized? And, what are the potential pitfalls of an AI-based approach for disaster risk analysis that we as a society must be cautious of?     

Risk Science Contributions: Three Illustrating Examples

This article aims to demonstrate that risk science is important for society, industry and all of us. Rather few people today, including scientists and managers, are familiar with what this science is about—its foundation and main features—and how it is used to gain knowledge and improve communication and decision making in real-life situations. The article seeks to meet this challenge, by presenting three examples, showing how risk science works to gain new generic, fundamental knowledge on risk concepts, principles, and methods, as well as supporting the practical tackling of actual risk problems.     

What's Wrong with Hazard‐Ranking Systems? An Expository Note

Two commonly recommended principles for allocating risk management resources to remediate uncertain hazards are: (1) select a subset to maximize risk-reduction benefits (e.g., maximize the von Neumann-Morgenstern expected utility of the selected risk-reducing activities), and (2) assign priorities to risk-reducing opportunities and then select activities from the top of the priority list down until no more can be afforded. When different activities create uncertain but correlated risk reductions, as is often the case in practice, then these principles are inconsistent: priority scoring and ranking fails to maximize risk-reduction benefits. Real-world risk priority scoring systems used in homeland security and terrorism risk assessment, environmental risk management, information system vulnerability rating, business risk matrices, and many other important applications do not exploit correlations among risk-reducing opportunities or optimally diversify risk-reducing investments. As a result, they generally make suboptimal risk management recommendations. Applying portfolio optimization methods instead of risk prioritization ranking, rating, or scoring methods can achieve greater risk-reduction value for resources spent.     

An Emerging New Risk Analysis Science: Foundations and Implications

To solve real‐life problems—such as those related to technology, health, security, or climate change—and make suitable decisions, risk is nearly always a main issue. Different types of sciences are often supporting the work, for example, statistics, natural sciences, and social sciences. Risk analysis approaches and methods are also commonly used, but risk analysis is not broadly accepted as a science in itself. A key problem is the lack of explanatory power and large uncertainties when assessing risk. This article presents an emerging new risk analysis science based on novel ideas and theories on risk analysis developed in recent years by the risk analysis community. It builds on a fundamental change in thinking, from the search for accurate predictions and risk estimates, to knowledge generation related to concepts, theories, frameworks, approaches, principles, methods, and models to understand, assess, characterize, communicate, and (in a broad sense) manage risk. Examples are used to illustrate the importance of this distinct/separate risk analysis science for solving risk problems, supporting science in general and other disciplines in particular.     

Foundational Challenges for Advancing the Field and Discipline of Risk Analysis

Risk analysis as a field and discipline is about concepts, principles, approaches, methods, and models for understanding, assessing, communicating, managing, and governing risk. The foundation of this field and discipline has been subject to continuous discussion since its origin some 40 years ago with the establishment of the Society for Risk Analysis and the Risk Analysis journal. This article provides a perspective on critical foundational challenges that this field and discipline faces today, for risk analysis to develop and have societal impact. Topics discussed include fundamental questions important for defining the risk field, discipline, and science; the multidisciplinary and interdisciplinary features of risk analysis; the interactions and dependencies with other sciences; terminology and fundamental principles; and current developments and trends, such as the use of artificial intelligence.     

Expert and Public Perception of Risk from Biotechnology

Risk perceptions of a series of biotechnology applications were examined in a public (nonexpert) sample and an expert sample. Compared with the experts, the public perceived all biotechnology applications as more risky. Both groups perceived food-related applications to be riskier than medical applications. Compared with the public, experts perceived both food and medical applications as less harmful and more useful. Experts also judged the risks posed from medical biotechnology applications as more familiar and acknowledged by people and science. Lay estimates of the risk of food applications were predicted by potential harm, potential benefits, science knowledge, and familiarity; experts' estimates were predicted only by harm and benefits. Lay estimates of the risk of medical applications were predicted by potential harm; experts' estimates were predicted by potential benefits, number and type of people exposed, and science knowledge. We discuss the implications of the results for risk communication about and management of different types of biotechnologies.     

Cases of real-life policies related to risk: How can they enhance risk analysis and risk science?

Policies on risk constitute a core topic of risk analysis and risk science, and it is common at risk conferences to present real-life cases of such policies, for example related to the handling of climate change and pandemics. Although these are of broad interest, showing how important issues in society are dealt with, it can be questioned to what extent and how these cases contribute to enhancing risk analysis and risk science. The present paper addresses this concern. It is argued that, in order to learn from the cases, they need in general to be more thoroughly followed up with discussions of concepts, principles, approaches, and methods for assessing, characterizing, communicating and handling risk. Describing a governmental policy on, for example, the handling of COVID-19 is a point of departure for interesting discussions concerning its justification and performance, in particular in relation to risk and the most updated knowledge from the risk analysis field. Such discussions are, however, often lacking. The paper points to some key obstacles and challenges for the learning process, including the difficulty of distinguishing between policies, policy analysis, and politics.     

An Interagency Comparison of Screening-Level Risk Assessment Approaches

Approaches to risk assessment have been shown to vary among regulatory agencies and across jurisdictional boundaries according to the different assumptions and justifications used. Approaches to screening-level risk assessment from six international agencies were applied to an urban case study focusing on benzo[a]pyrene (B[a]P) exposure and compared in order to provide insight into the differences between agency methods, assumptions, and justifications. Exposure estimates ranged four-fold, with most of the dose stemming from exposure to animal products (8-73%) and plant products (24-88%). Total cancer risk across agencies varied by two orders of magnitude, with exposure to air and plant and animal products contributing most to total cancer risk, while the air contribution showed the greatest variability (1-99%). Variability in cancer risk of 100-fold was attributed to choices of toxicological reference values (TRVs), either based on a combination of epidemiological and animal data, or on animal data. The contribution and importance of the urban exposure pathway for cancer risk varied according to the TRV and, ultimately, according to differences in risk assessment assumptions and guidance. While all agency risk assessment methods are predicated on science, the study results suggest that the largest impact on the differential assessment of risk by international agencies comes from policy and judgment, rather than science.     

In Search of Perfect Foresight? Policy Bias,Management of Unknowns,and What Has Changed in Science Policy Since the Tohoku Disaster

The failure to foresee the catastrophic earthquakes, tsunamis, and nuclear accident of 2011 has been perceived by many in Japan as a fundamental shortcoming of modern disaster risk science. Hampered by a variety of cognitive and institutional biases, the conventional disaster risk management planning based on the “known risks” led to the cascading failures of the interlinked disaster risk management (DRM) apparatus. This realization led to a major rethinking in the use of science for policy and the incorporations of lessons learned in the country's new DRM policy. This study reviews publicly available documents on expert committee discussions and scientific articles to identify what continuities and changes have been made in the use of scientific knowledge in Japanese risk management. In general, the prior influence of cognitive bias (e.g., overreliance on documented hazard risks) has been largely recognized, and increased attention is now being paid to the incorporation of less documented but known risks. This has led to upward adjustments in estimated damages from future risks and recognition of the need for further strengthening of DRM policy. At the same time, there remains significant continuity in the way scientific knowledge is perceived to provide sufficient and justifiable grounds for the development and implementation of DRM policy. The emphasis on “evidence‐based policy” in earthquake and tsunami risk reduction measures continues, despite the critical reflections of a group of scientists who advocate for a major rethinking of the country's science‐policy institution respecting the limitations of the current state science.     

Risk analysis under attack: How risk science can address the legal,social, and reputational liabilities faced by risk analysts

The role of the risk analyst is critical in understanding and managing uncertainty. However, there is another type of uncertainty that is rarely discussed: The legal, social, and reputational liabilities of the risk analyst. Recent events have shown that professionals participating in risk analysis can be held personally liable. It is timely and important to ask: How can risk science guide risk analysis with consideration of those liabilities, particularly in response to emerging and unprecedented risk. This paper studies this topic by: (1) Categorizing how professionals with risk analysis responsibilities have historically been held liable, and (2) developing a framework to address uncertainty related to those potential liabilities. The result of this framework will enable individual analysts and organizations to investigate and manage the expectations of risk analysts and others as they apply risk principles and methods. This paper will be of interest to risk researchers, risk professionals, and industry professionals who seek maturity within their risk programs.     

Some Considerations on the Definition of Risk Based on Concepts of Systems Theory and Probability

The concept of risk has been applied in many modern science and technology fields. Despite its successes in many applicative fields, there is still not a well‐established vision and universally accepted definition of the principles and fundamental concepts of the risk assessment discipline. As emphasized recently, the risk fields suffer from a lack of clarity on their scientific bases that can define, in a unique theoretical framework, the general concepts in the different areas of application. The aim of this article is to make suggestions for another perspective of risk definition that could be applied and, in a certain sense, generalize some of the previously known definitions (at least in the fields of technical and scientific applications). By drawing on my experience of risk assessment in different applicative situations (particularly in the risk estimation for major industrial accidents, and in the health and ecological risk assessment for contaminated sites), I would like to revise some general and foundational concepts of risk analysis in as consistent a manner as possible from the axiomatic/deductive point of view. My proposal is based on the fundamental concepts of the systems theory and of the probability. In this way, I try to frame, in a single, broad, and general theoretical context some fundamental concepts and principles applicable in many different fields of risk assessment. I hope that this article will contribute to the revitalization and stimulation of useful discussions and new insights into the key issues and theoretical foundations of risk assessment disciplines.     

Comparative Risk and Policy Analysis in Environmental Health

There is increasing interest in the integration of quantitative risk analysis with benefit-cost and cost-effectiveness methods to evaluate environmental health policy making and perform comparative analyses. However, the combined use of these methods has revealed deficiencies in the available methods, and the lack of useful analytical frameworks currently constrains the utility of comparative risk and policy analyses. A principal issue in integrating risk and economic analysis is the lack of common performance metrics, particularly when conducting comparative analyses of regulations with disparate health endpoints (e.g., cancer and noncancer effects or risk-benefit analysis) and quantitative estimation of cumulative risk, whether from exposure to single agents with multiple health impacts or from exposure to mixtures. We propose a general quantitative framework and examine assumptions required for performing analyses of health risks and policies. We review existing and proposed risk and health-impact metrics for evaluating policies designed to protect public health from environmental exposures, and identify their strengths and weaknesses with respect to their use in a general comparative risk and policy analysis framework. Case studies are presented to demonstrate applications of this framework with risk-benefit and air pollution risk analyses. Through this analysis, we hope to generate discussions regarding the data requirements, analytical approaches, and assumptions required for general models to be used in comparative risk and policy analysis.     

Engaging Expert Peers in the Development of Risk Assessments

The participation of external technical experts in the development of risk assessment documents and methodologies has expanded and evolved in recent years. Many government agencies and authoritative organizations have experts peer review important works to evaluate the scientific and technical defensibility and judge the strength of the assumptions and conclusions (OMB, 2004; IPCS, 2005; IARC, 2006; Health Canada, 2007; U.S. EPA, 2006). Expert advice has been solicited in other forms of peer involvement, including peer consultation in, for example, the U.S. EPA's Voluntary Children's Chemical Evaluation Program (VCCEP). This article discusses how the principles and practices of peer review can be extended to other types of peer involvement activities (i.e., peer input and peer consultation) to develop high-quality risk assessment work products. A comprehensive process for incorporating peer input, peer consultation, and peer review into risk assessment science is outlined. Four key principles for peer involvement-independence, inclusion of appropriate experts, transparency, and a robust scientific process-are discussed. Recent examples of peer involvement in the development of Health Canada's Priority Substances and Domestic Substance List (DSL) programs under the Canadian Environmental Protection Act (CEPA) serve to highlight the concepts.     

Workforce/Population,Economy, Infrastructure,Geography, Hierarchy,and Time (WEIGHT): Reflections on the Plural Dimensions of Disaster Resilience

The concept of resilience and its relevance to disaster risk management has increasingly gained attention in recent years. It is common for risk and resilience studies to model system recovery by analyzing a single or aggregated measure of performance, such as economic output or system functionality. However, the history of past disasters and recent risk literature suggest that a single-dimension view of relevant systems is not only insufficient, but can compromise the ability to manage risk for these systems. In this article, we explore how multiple dimensions influence the ability for complex systems to function and effectively recover after a disaster. In particular, we compile evidence from the many competing resilience perspectives to identify the most critical resilience dimensions across several academic disciplines, applications, and disaster events. The findings demonstrate the need for a conceptual framework that decomposes resilience into six primary dimensions: workforce/population, economy, infrastructure, geography, hierarchy, and time (WEIGHT). These dimensions are not typically addressed holistically in the literature; often they are either modeled independently or in piecemeal combinations. The current research is the first to provide a comprehensive discussion of each resilience dimension and discuss how these dimensions can be integrated into a cohesive framework, suggesting that no single dimension is sufficient for a holistic analysis of a disaster risk management. Through this article, we also aim to spark discussions among researchers and policymakers to develop a multicriteria decision framework for evaluating the efficacy of resilience strategies. Furthermore, the WEIGHT dimensions may also be used to motivate the generation of new approaches for data analytics of resilience-related knowledge bases.     

How effective are community-based disaster reduction strategies? Evidence from the largest-scale program so far

Strategies of community-based disaster risk reduction have been advocated for more than 2 decades. However, we still lack in-depth quantitative assessments of the effectiveness of such strategies. Our research is based on a national experiment in this domain: the “Comprehensive Disaster Reduction Demonstration Community” project, a governmental program running in China since 2007. Information on more than 11,000 demonstration communities was collected. Combined with the local disaster information and socioeconomic conditions, the spatiotemporal characteristics of these communities over 12 years and their differences in performance by region and income group were analyzed. We performed an attribution analysis for disaster risk reduction effectiveness. This is the first time a series of quantitative evaluation methods have been applied to verify the effectiveness of a large-scale community-based disaster risk reduction project, both from the perspective of demonstrative effects and loss reduction benefits. Here, we find that the project is obviously effective from these two perspectives, and the disaster loss reduction effectiveness illustrates clear regional differences, where the regional economic level and hazard severity act as important drivers. Significant differences of urban-rural and income call for matching fortification measures, and the dynamic management of demonstration community size is required, since the loss reduction benefit converges when the penetration rate of the demonstration community reaches approximately 4% in a province. These and further results provide diverse implications for community-based disaster risk reduction policies and practices.     

Why income inequality indexes do not apply to health risks

Several recent papers have sought to apply inequality measures from economics, such as the Atkinson Index (AI) for inequality of income distributions, to compare the risk inequality of different mortality risk distributions in an effort to help promote efficiency and environmental justice in pollution-reducing interventions. Closer analysis suggests that such applications are neither logically coherent nor necessarily ethically desirable. Risk inequality comparisons should be based on axioms that apply to probabilistic risks, and should consider the multidimensional and time-varying nature of individual and community risks in order to increase efficiency and justice over time and generations. In light of the limitations of the AI applied to mortality risk distributions, it has not been demonstrated to have ethical or practical value in helping policymakers to identify air pollution management interventions that reduce (or minimize) risk and risk inequity.     

Foundational Issues in Risk Assessment and Risk Management

In spite of the maturity reached by many of the methods used in risk assessment and risk management, broad consensus has not been established on fundamental concepts and principles. The risk fields still suffer from a lack of clarity on many key scientific pillars. The purpose of this article is to point to this situation and through some illustrating examples discuss the challenges that the fields here face. Moreover, the purpose of the article is to reflect on how to improve the present situation and enhance the risk fields. We argue that the establishment of some common scientific pillars as well as a strong and continuous research focus on foundational issues are critical success factors. The article specifically addresses the role of the peer‐reviewed journals and the international standards in the fields. We hope that the article can contribute to a revitalization of the discussion of foundational issues in risk assessment and risk management.     

A risk science perspective on liability/guilt and uncertainty judgements in courts

This article aims to provide new insights about risk and uncertainty in law contexts, by incorporating ideas and principles of contemporary risk science. The main focus is on one particular aspect of the law: its operation in courts where a defendant has been charged with a violation of civil or criminal law. Judgements about risk and uncertainty—typically using the probability concept—and how these relate to the evidence play a central role in such situations. The decision on whether the defendant is liable/guilty or not may strongly depend on how these concepts are understood and communicated. Considerable work has been conducted to provide theoretical and practical foundations for the risk and uncertainty characterizations in these contexts. Yet, it can be argued that a proper foundation for linking the evidence and the uncertainty (probability) judgements is lacking, the result being poor communication in courts about risk and uncertainties. The present article seeks to clarify what the problems are and provide guidance on how to rectify them.     

Maximizing Health Benefits and Minimizing Inequality: Incorporating Local‐Scale Data in the Design and Evaluation of Air Quality Policies

The U.S. Environmental Protection Agency undertook a case study in the Detroit metropolitan area to test the viability of a new multipollutant risk‐based (MP/RB) approach to air quality management, informed by spatially resolved air quality, population, and baseline health data. The case study demonstrated that the MP/RB approach approximately doubled the human health benefits achieved by the traditional approach while increasing cost less than 20%—moving closer to the objective of Executive Order 12866 to maximize net benefits. Less well understood is how the distribution of health benefits from the MP/RB and traditional strategies affect the existing inequalities in air‐pollution‐related risks in Detroit. In this article, we identify Detroit populations that may be both most susceptible to air pollution health impacts (based on local‐scale baseline health data) and most vulnerable to air pollution (based on fine‐scale PM_2.5 air quality modeling and socioeconomic characteristics). Using these susceptible/vulnerable subpopulation profiles, we assess the relative impacts of each control strategy on risk inequality, applying the Atkinson Index (AI) to quantify health risk inequality at baseline and with either risk management approach. We find that the MP/RB approach delivers greater air quality improvements among these subpopulations while also generating substantial benefits among lower‐risk populations. Applying the AI, we confirm that the MP/RB strategy yields less PM_2.5 mortality and asthma hospitalization risk inequality than the traditional approach. We demonstrate the value of this approach to policymakers as they develop cost‐effective air quality management plans that maximize risk reduction while minimizing health inequality.