What do we loose when machines take the decisions?

This paper concerns the technical issues raised when humans are replaced by artificial intelligence (AI) in organisational decision making, or decision making in general. Such automation of human tasks and decision making can of course be beneficial through saving human resources, and through (ideally) leading to better solutions and decisions. However, to guarantee better decisions, the current AI techniques still have some way to go in most areas, and many of the techniques also suffer from weaknesses such as lack of transparency and explainability. The goal of the paper is not to argue against using any kind of AI in organisational decision making. AI techniques have a lot to offer, and can for instance assess a lot more possible decisions—and much faster—than any human can. The purpose is just to point to the weaknesses that AI techniques still have, and that one should be aware of when considering to implement AI to automate human decisions. Significant current AI research goes into reducing its limitations and weaknesses, but this is likely to become a fairly long-term effort. People and organisations might be tempted to fully automate certain crucial aspects of decision making without waiting for these limitations and weaknesses to be reduced—or, even worse, not even being aware of those weaknesses and what is lost in the automatisation process.

     

Cognitive and Motivational Biases in Decision and Risk Analysis

Behavioral decision research has demonstrated that judgments and decisions of ordinary people and experts are subject to numerous biases. Decision and risk analysis were designed to improve judgments and decisions and to overcome many of these biases. However, when eliciting model components and parameters from decisionmakers or experts, analysts often face the very biases they are trying to help overcome. When these inputs are biased they can seriously reduce the quality of the model and resulting analysis. Some of these biases are due to faulty cognitive processes; some are due to motivations for preferred analysis outcomes. This article identifies the cognitive and motivational biases that are relevant for decision and risk analysis because they can distort analysis inputs and are difficult to correct. We also review and provide guidance about the existing debiasing techniques to overcome these biases. In addition, we describe some biases that are less relevant because they can be corrected by using logic or decomposing the elicitation task. We conclude the article with an agenda for future research.     

TOWARD UNDERSTANDING THE CONTENTS OF THE “BLACK BOX” FOR PREDICTING COMPLEX DECISION-MAKING OUTCOMES

The purpose of this research is to show the usefulness of three relatively simple nonlinear classification techniques for policy-capturing research where linear models have typically been used. This study uses 480 cases to assess the decision-making process used by 24 experienced national bank examiners in classifying commercial loans as acceptable or questionable. The results from multiple discriminant analysis (a linear technique) are compared to those of chi-squared automatic interaction detector analysis (a search technique), log-linear analysis, and logit analysis. Results show that while the four techniques are equally accurate in predicting loan classification, chi-squared automatic interaction detector analysis (CHAID) and log-linear analysis enable the researcher to analyze the decision-making structure and examine the “human” variable within the decision-making process. Consequently, if the sole purpose of research is to predict the decision maker's decisions, then any one of the four techniques turns out to be equally useful. If, however, the purpose is to analyze the decision-making process as well as to predict decisions, then CHAID or log-linear techniques are more useful than linear model techniques.     

The Relationship between Procedural Rationality and Political Behavior in Strategic Decision Making

Despite the central place of rationality and political behavior in the decision-making literature, we know little about the relationship between these two dimensions. Can decisions be made using both rational and political methods, or must managers use one approach or the other? These questions were addressed in a study of 61 strategic decisions in 24 companies using a multiple-informant, structured interview protocol. Results indicate that procedural rationality and political behavior are independent dimensions of the strategic decision-making process. The implications of our findings for future strategic decision-making research are discussed.     

供应链中延迟研究的发展与展望

本文从延迟的定义与分类、不同延迟策略的建模分析、延迟的实施研究等几个方面综述供应链中延迟研究的主要成果.认为供应链延迟的进一步研究可以结合顾客化定制、订单完成周期与顾客行为相关的决策展开.     

A framework of decision making for technological change

The first article in this series2 drew a distinction between the costs of line operations- or current outputs—and those costs devoted to investment, or future outputs. As investment encompasses technological change—in practice it is dominated by it—this article examines some of the broader issues underlying decisions to invest and describes in outline a study being made of the economic criteria that are relevant to such decisions. In the longer term the objective is to develop a procedure or framework of criteria to assist the process of decision making; and this would include a frame of reference—a set of statistical time-series and analyses designed for the purpose.     

A Comparative Analysis of PRA and Intelligent Adversary Methods for Counterterrorism Risk Management

In counterterrorism risk management decisions, the analyst can choose to represent terrorist decisions as defender uncertainties or as attacker decisions. We perform a comparative analysis of probabilistic risk analysis (PRA) methods including event trees, influence diagrams, Bayesian networks, decision trees, game theory, and combined methods on the same illustrative examples (container screening for radiological materials) to get insights into the significant differences in assumptions and results. A key tenent of PRA and decision analysis is the use of subjective probability to assess the likelihood of possible outcomes. For each technique, we compare the assumptions, probability assessment requirements, risk levels, and potential insights for risk managers. We find that assessing the distribution of potential attacker decisions is a complex judgment task, particularly considering the adaptation of the attacker to defender decisions. Intelligent adversary risk analysis and adversarial risk analysis are extensions of decision analysis and sequential game theory that help to decompose such judgments. These techniques explicitly show the adaptation of the attacker and the resulting shift in risk based on defender decisions.     

Intergenerational Risk Decision Making: A Practical Example

There is no such thing as intergenerational decision making, at least not yet. In fact, there is no such thing as intragenerational decision making in the context of maximizing overall social good given resource limitations, there are just decisions being made in an ad hoc fashion. Even if one assumes that there is such a thing as intragenerational decision making, no uniform standard or guidance exists to make societal decisions for the common good. Risks to society are judged unevenly within the same agency and across agencies. Decisions are made in isolation and not weighed in the societal context of what is intra or intergenerationally important. The National Academy of Public Administration (NAPA) has set forth a framework for intergenerational decision making that provides a consistent and fair basis for making tough decisions in order to address difficult issues such as the long-term disposal of nuclear wastes. NAPA recognizes that there is an intergenerational obligation that must encompass broader questions than the narrow issue of waste disposal since resources are finite and needs are great. The fundamental principles are based on sustainability with the overarching objective that "no generation should needlessly, now or in the future, deprive its successors of the opportunity to enjoy a quality of life equivalent to its own." Coupled with this objective are four supporting principles of trusteeship, sustainability, chain of obligation, and precaution. The NAPA process also recognizes that no decision can be final and that a "rolling future" view is better than making decisions for "all time." It attempts to balance the needs of the present with those of the future in an open and transparent process that is aimed at producing a decision, not just endless analysis. The U.S. Congress and president should develop a rational standard by which to judge laws that involve intra and intergenerational issues relative to the overall societal good. Present regulations need to be evaluated relative to a uniform level of risk and benefit to assess where the limited money available can do the most good for both the present and future generations in the context of NAPA sustainability principles. It is hoped that decision makers will take a serious look at this process since it can work to resolve stakeholder stalemate.     

Group Selling,Product Durability,and Consumer Behavior

Firms producing complementary goods often strategically form groups and jointly sell their products to better coordinate their decisions. For consumer durables, decisions about such collaboration might be complicated due to two factors. Because of their durability and presence of used goods markets, such products engender “future” price competition between new and used goods. On the other hand, consumers of such products might be forward‐looking and patient, both of which affect their purchasing behavior. In this study, we study how the above product and consumer characteristics interact to affect the group selling decisions of complementary firms. We do so through a two‐period model consisting of a value chain with two upstream manufacturers and a downstream retailer. When consumers are relatively impatient and reluctant to wait to buy later, group selling by manufacturers will take place only when the end product is relatively perishable, that is, product durability is low. However, if consumers are patient, that is, willing to wait, collaboration happens only when the end product is quite durable; for relatively perishable products the manufacturers sell their products separately. We also comment on how our results are affected by factors like manufacturers directly selling to end consumers or there being multiple opportunities to decide whether or not to use group selling strategy.     

Is there potential in assessing for high-potential? Evaluating the relationships between performance ratings,leadership assessment data,designated high-potential status and promotion outcomes in a global organization

While leadership is a critical component of organizational success, organizations need to be more effective at classifying talent based on future potential vs current performance. This need raises the question as to whether the assessment of leadership potential is truly adding value or enhancing the talent review and classification process. Data from 9,784 participants in PepsiCo’s Leadership Assessment and Development (LeAD) program—a multi-trait, multi-method (MTMM) assessment and development process—are used to examine relationships between individual performance, assessed potential, organizationally designated potential, and promotion rates post assessment. Results from the analyses indicate that assessed potential provides unique variance above performance alone in determining process outcomes (i.e., designated potential and promotions), and therefore can be useful for making more informed talent management decisions. The applications of these measures within the current organization as well as limitations of the study and areas for future research are discussed.     

Stimulating Early Adoption of New Products through Channel Disintegration

Conventional wisdom holds that adding layers to a distribution channel is detrimental to the interests of consumers and the channel that serves them. In contrast, our study indicates that a disintegrated channel structure can be desirable in some instances. When consumers have valuation uncertainty prior to consuming a product, having an independent retailer may boost both channel profits and consumer surplus relative to direct selling by an integrated firm. The quandary in selling such products is that after early adopters make their purchase decisions, the seller may alter prices in such a way that makes early adopters' decisions appear suboptimal in hindsight. Since the seller cannot credibly commit to future prices, customers are reluctant to adopt early, choosing instead to delay their purchase decisions. This delay is certainly detrimental to the interest of the distribution channel, but the rejection of the early adoption discount can equally reduce consumer surplus. This problem can be mitigated by introducing an independent retailer. The familiar double marginalization “problem” from channel disintegration can credibly assure customers of unfavorable future prices for late adoption. This assurance attracts more customers to seek early adoption, leading to lower overall retail prices, increased supply, and higher consumer and producer surpluses.     

The planning function and input-output analysis

Long range planning has been defined as a decision-making process involving the commitment of resources—money, people, time, capital—today, the payback or return on which will not be realized until some future period. Since the heart of the planning process is the making of decisions about allocations in some systematic way, economic analysis, armed with its concepts of opportunity cost, indifference theory, marginal analysis, and investment theory, provides much of the theoretical framework for good long range planning models.

If planning is to be done in a scientific manner, there shouldexist common principles and study approaches, regardless of the nature of the organism being planned for. Thus planning for the individual, family, company, industry, association or state should vary according to the relevant information to be used in making decisions and the priorities of the objectives but not in the essential method by which decisions are arrived at. This is major theme of the paper.

As a result of more, than a decade of experience as a ‘planner’ for a number of diverse business and government groups, and through the process of articulating this planning experience as a teacher, I have developed a systematic approach to the planning process. The model has its own conceptual framework that includes, as information filters and display devices, unique adaptations of input-output notation. First in this article I will highlight some of the elements of the planning process, as I have constructed them. Then I will proceed to describe the specifications of the planning system, with particular emphasis on the use of input-output analysis.     

Consortium for Atlantic Regional Assessment: Information Tools for Community Adaptation to Changes in Climate or Land Use

To inform local and regional decisions about protecting short-term and long-term quality of life, the Consortium for Atlantic Regional Assessment (CARA) provides data and tools (for the northeastern United States) that can help decision makers understand how outcomes of their decisions could be affected by potential changes in both climate and land use. On an interactive, user-friendly website, CARA has amassed data on climate (historical records and future projections for seven global climate models), land cover, and socioeconomic and environmental variables, along with tools to help decision makers tailor the data for their own decision types and locations. CARA Advisory Council stakeholders help identify what information and tools stakeholders would find most useful and how to present these; they also provide in-depth feedback for subregion case studies. General lessons include: (1) decision makers want detailed local projections for periods short enough to account for extreme events, in contrast to the broader spatial and temporal observations and projections that are available or consistent at a regional level; (2) stakeholders will not use such a website unless it is visually appealing and easy to find the information they want; (3) some stakeholders need background while others want to go immediately to data, and some want maps while others want text or tables. This article also compares what has been learned across case studies of Cape May County, New Jersey, Cape Cod, Massachusetts, and Hampton Roads, Virginia, relating specifically to sea-level rise. Lessons include: (1) groups can be affected differently by physical dangers compared with economic dangers; (2) decisions will differ according to decision makers' preferences about waiting and risk tolerance; (3) future scenarios and maps can help assess the impacts of dangers to emergency evacuation routes, homes, and infrastructure, and the natural environment; (4) residents' and decision makers' perceptions are affected by information about potential local impacts from global climate change.     

MANAGING PRODUCT VARIETY: AN INTEGRATIVE REVIEW AND RESEARCH DIRECTIONS

Variety management has emerged as a crucial dimension of successful business practice. In this paper, I first provide a framework for managerial decisions about variety. Variety‐creation decisions determine the amount, type, and timing of end‐product variety, while variety‐implementation decisions focus on the design and operation of internal processes and a supply chain to support a firm's variety‐creation strategy. I organize variety‐related decisions into four key decision themes in variety creation: 1) dimensions of variety, 2) product architecture, 3) degree of customization, and 4) timing; and three key decision themes in variety implementation: 1) process capabilities, 2) points of variegation, and 3) day‐to‐day decisions. I describe each theme and review the relevant literature on each theme, with a focus on research that provides insight to problems faced in practice. Finally, I identify untapped avenues for future research that would be of value to the practicing manager, paying special attention to interdependencies among decision themes.     

Cybernetic approach to investment decision making

This paper summarizes the results of our research into applications of cybernetic concepts and artificial intelligence techniques to investment analysis: it outlines the philosophy underlying the cybernetic approach to market forecasting and investment selection. This approach is computer oriented—it enables us to automate or program directly the complex judgmental aspects of investment decision making. The cybernetic approach alleviates some deficiencies of conventional statistical methods. Specifically, it (1) explicitly includes the time dimension into investment analysis, (2) is based on methods for decision making under uncertainty rather than risk, (3) yields computationally feasible methods for coping with the high complexity in investment analysis, and (4) yields decisions that are optimal rather than satisficing, i.e. performance of the Cybernetic Investment Decision System (CIDS) gradually improves during its operation through learning from past experience. A simplified CIDS has been implemented and tested in actual investment analysis. The experimental results of these tests indicate that through the cybernetic approach quality of investment decisions can be improved.     

Robustness of linear models in dynamic multivariate predictions

Linear aggregation models employing unit and equal weights have been shown to be superior to human decisions in a surprising range of decision situations. In addition, decisions based on these models have often been found to be superior to those based on linear regression models (LRMs). This general issue was explored for repetitive decisions in production planning. The problem considered differs in several aspects from the types of problems investigated previously: (1) the problem is dynamic rather than static; (2) a set (or vector) of interactive decisions dependent on previous decisions is required to be made, where a decision in stage t, the dependent variable, becomes an independent variable in stage t + 1; and (3) the criterion function is cost with a quadratic loss function (rather than the correlation measure of R²). Moreover, since repetitive decisions were involved, the parameters of the models were estimated using past human decisions. These were used to predict specific values of the decision variables (rather than rank order), which in turn were employed recursively to predict values of the decision variables at subsequent stages. While decisions from equal weighting rules were found to be superior to human decisions and not greatly inferior to decisions from linear regression models, decisions from unit weighting rules performed poorly. The rationale for such performance is discussed, indicating that previous theoretical and empirical research on linear weighting models is not generally applicable to dynamic, multivariate interactive decisions problems with lagged variables.     

A model for multinational corporate money management

A centralized multiple-period money management model is developed and applied to the case of a simulated US based MNC with subsidiaries in the United Kingdom, Canada and West Germany. Only the first of three period's decisions are intended for implementation in the moving horizon model. The inclusion of two future periods permits current actions to be executed with imperfect but valuable information contained in forecasts of future spot exchange rates produced by moving-horizon least-squares generators. For the set of assumptions invoked it is seen that the centralized model provides significantly better profit results than zero-net exposure approaches.     

The Fairness of Decision Rules

Prior work on resource allocation has generally considered only a small number of allocation rules, usually reflecting equity or equality. We use a scenario study to examine the effect of eight different allocation rules (past performance, future performance, rank, random draw, chance meetings, business need, personal need, and political reasons) on recipient reactions to the gain or loss of three different kinds of resources in an organizational setting. We find evidence that allocations based on past performance and random draw rules lead to the highest fairness perceptions and the lowest expectations that the decisions made will lead to intragroup conflict. However, fairness judgments are also influenced by a variety of other factors, such as the type of resource being allocated and whether the recipient is advantaged or disadvantaged relative to others in the workgroup (what we term the “egocentric interaction”). We discuss how our results might influence managers’ allocation decisions.     

A generic accounting model to support operations management decisions

Information systems are generally unable to generate information about the financial consequences of operations management decisions. This is because the procedures for determining the relevant accounting information for decision support are not formalised in ways that can be implemented in information systems. This paper describes a formalised procedure, which is based on the following theoretical propositions: (i) cost behaviour is described on the basis of a company's contracts for purchasing and selling resources, and (ii) hierarchical relationships between decisions are recognised, because some decisions have to be made earlier than others. Earlier decisions determine the feasible alternatives for later decisions (e.g. through constraints in available production capacity or components), and the plans that supported the earlier decisions serve as instructions for later decisions. The procedure can be implemented in information systems to provide accounting information in case later decisions deviate from these instruc tions (within the limits of the real-world constraints) because new information becomes available.