Search

Objective We explore users’ and observers’ subjective assessments of human and automation capabilities and human causal responsibility for outcomes. Background In intelligent systems and advanced automation, human responsibility for outcomes becomes equivocal, as do subjective perceptions of responsibility. In particular, actors who actively work with a system may perceive responsibility differently from observers. Method In a laboratory experiment with pairs of participants, one participant (the “actor”) performed a decision task, aided by an automated system, and the other (the “observer”) passively observed the actor. We compared the perceptions of responsibility between the two roles when interacting with two systems with different capabilities. Results Actors’ behavior matched the theoretical predictions, and actors and observers assessed the system and human capabilities and the comparative human responsibility similarly. However, actors tended to relate adverse outcomes more to system characteristics than to their own limitations, whereas the observers insufficiently considered system capabilities when evaluating the actors’ comparative responsibility. Conclusion When intelligent systems greatly exceed human capabilities, users may correctly feel they contribute little to system performance. They may interfere more than necessary, impairing the overall performance. Outside observers, such as managers, may overweigh users’ contribution to outcomes, holding users responsible for adverse outcomes when they rightly trusted the system. Application Presenting users of intelligent systems and others with performance measures and the comparative human responsibility may help them calibrate subjective assessments of performance, reducing users’ and outside observers’ biases and attribution errors.

Intelligent systems and advanced automation are involved in information collection and evaluation, decision-making, and the implementation of chosen actions. In such systems, human responsibility becomes equivocal. Understanding human causal responsibility is particularly important when systems can harm people, as with autonomous vehicles or, most notably, with autonomous weapon systems (AWSs). Using information theory, we developed a responsibility quantification (ResQu) model of human causal responsibility in intelligent systems and demonstrated its applications on decisions regarding AWS. The analysis reveals that comparative human responsibility for outcomes is often low, even when major functions are allocated to the human. Thus, broadly stated policies of keeping humans in the loop and having meaningful human control are misleading and cannot truly direct decisions on how to involve humans in advanced automation. The current model assumes stationarity, full knowledge regarding the characteristic of the human and automation, and ignores temporal aspects. It is an initial step toward the development of a comprehensive responsibility model that will make it possible to quantify human causal responsibility. The model can serve as an additional tool in the analysis of system design alternatives and policy decisions regarding human causal responsibility, providing a novel, quantitative perspective on these matters. Note to Practitioners —We developed a theoretical model and a quantitative measure for computing the comparative human causal responsibility in the interaction with intelligent systems and advanced automation. Our responsibility measure can be applied by practitioners (system designers, regulators, and so on) to estimate user responsibility in specific system configurations. This can serve as an additional tool in the comparison between alternative system designs or deployment policies, by relating different automation design options to their predicted effect on the users’ responsibility. To apply the model (which is based on entropy and mutual information) to real-world systems, one must deduce the underlying distributions, either from known system properties or from empirical observations, taken over time. The initial version of the model we present here assumes that the combined human–automation system is stationary and ergodic. Real-world systems may not be stationary and ergodic or cannot be observed sufficiently to allow accurate estimates of the required input of multivariate probabilities, in which case the computed responsibility values should be treated with caution. Nevertheless, the construction of a ResQu information flow model, combined with sensitivity analyses of how changes in the input probabilities and assumptions affect the responsibility measure, will often reveal important qualitative properties and supply valuable insights regarding the general level of meaningful human involvement and comparative responsibility in a system.

The “intuitive” trust people feel when encountering robots in public spaces is a key determinant of their willingness to cooperate with these robots. We conducted four experiments to study this topic in the context of peacekeeping robots. Participants viewed scenarios, presented as static images or animations, involving a robot or a human guard performing an access-control task. The guards interacted more or less politely with younger and older male and female people. Our results show strong effects of the guard's politeness. The age and sex of the people interacting with the guard had no significant effect on participants' impressions of its attributes. There were no differences between responses to robot and human guards. This study advances the notion that politeness is a crucial determinant of people's perception of peacekeeping robots.

Next-generation multi-specific antibody therapeutics (MSATs) are engineered to combine several functional activities into one molecule to provide higher efficacy compared to conventional, mono-specific antibody therapeutics. However, highly engineered MSATs frequently display poor yields and less favorable drug-like properties (DLPs), which can adversely affect their development. Systematic screening of a large panel of MSAT variants in very high throughput (HT) is thus critical to identify potent molecule candidates with good yield and DLPs early in the discovery process. Here we report on the establishment of a novel, format-agnostic platform process for the fast generation and multiparametric screening of tens of thousands of MSAT variants. To this end, we have introduced full automation across the entire value chain for MSAT engineering. Specifically, we have automated the in-silico design of very large MSAT panels such that it reflects precisely the wet-lab processes for MSAT DNA library generation. This includes mass saturation mutagenesis or bulk modular cloning technologies while, concomitantly, enabling library deconvolution approaches using HT Sanger DNA sequencing. These DNA workflows are tightly linked to fully automated downstream processes for compartmentalized mammalian cell transfection expression, and screening of multiple parameters. All sub-processes are seamlessly integrated with tailored workflow supporting bioinformatics. As described here, we used this platform to perform multifactor optimization of a next-generation bispecific, cross-over dual variable domain-Ig (CODV-Ig). Screening of more than 25,000 individual protein variants in mono- and bispecific format led to the identification of CODV-Ig variants with over 1,000-fold increased potency and significantly optimized production titers, demonstrating the power and versatility of the platform.

Binary decision aids, such as alerts, are a simple and widely used form of automation. The formal analysis of a user's task performance with an aid sees the process as the combination of information from two detectors who both receive input about an event and evaluate it. The user's decisions are based on the output of the aid and on the information, the user obtains independently. We present a simple method for computing the maximal benefits a user can derive from a binary aid as a function of the user's and the aid's sensitivities. Combining the user and the aid often adds little to the performance the better detector could achieve alone. Also, if users assign non-optimal weights to the aid, performance may drop dramatically. Thus, the introduction of a valid aid can actually lower detection performance, compared to a more sensitive user working alone. Similarly, adding a user to a system with high sensitivity may lower its performance. System designers need to consider the potential adverse effects of introducing users or aids into systems., Proceedings, 2nd IEEE International Conference on Human-Machine Systems (2021)

Cyber attacks and the associated costs made cybersecurity a vital part of any system. User behavior and decisions are still a major part in the coping with these risks. We developed a model of optimal investment and human decisions with security measures, given that the effectiveness of each measure depends partly on the performance of the others. In an online experiment, participants classified events as malicious or non-malicious, based on the value of an observed variable. Prior to making the decisions, they had invested in three security measures - a firewall, an IDS or insurance. In three experimental conditions, maximal investment in only one of the measures was optimal, while in a fourth condition, participants should not have invested in any of the measures. A previous paper presents the analysis of the investment decisions. This paper reports users’ classifications of events when interacting with these systems. The use of security mechanisms helped participants gain higher scores. Participants benefited in particular from purchasing IDS and/or Cyber Insurance. Participants also showed higher sensitivity and compliance with the alerting system when they could benefit from investing in the IDS. Participants, however, did not adjust their behavior optimally to the security settings they had chosen. The results demonstrate the complex nature of risk-related behaviors and the need to consider human abilities and biases when designing cyber security systems.

Image processing has high computational requirements. As many papers (including [Frisk 2010; N. Zhang and Wang 2010; Yang et al. 2008]) have indicated, a number of image processing operations can be optimized by heavy parallelization of the computation. Currently, one of the best options for parallelized image processing is using general purpose computation on graphics processing units (GPGPU).In the area of GPGPU, a rather wide range of APIs is available and finding an appropriate choice for a project is challenging. In the course of being able to provide guidance on selecting the right API in an image processing context, four GPU programming models were compared according to their platform independence, usability, and performance. The four investigated APIs are CUDA, OpenCL 1.2, Vulkan, and SYCL.To gain information on the metrics usability and performance a test project was created, implementing a number of image processing tasks using the four GPU programming models. The test project was designed as a library, where the 4 investigated APIs could be replaced at compile-time and new operations could be easily added. The implemented reference algorithms form a pipeline for processing images from polarization cameras. The targeted platforms were mid- to high-end desktop PCs as well as embedded single board platforms ranging from Odroid to Nvidia Jetson. On all devices, either MSVC 19.x or GCC 7+ were used as host compilers. CUDA tests were compiled with NVCC and SYCL was represented by ComputeCpp and hipSYCL depending on the device.The direct comparison of the APIs, done for this work, cumulates into a decision matrix to aid organizations in their process of selecting an offloading API. It shows, for example, that CUDA and SYCL are comparable in terms of development cost, support of modern C++ and ease of integration in existing code bases. When considering modern standard adoption and GPU portability Vulkan has advantages, while OpenCL and SYCL can run on CPUs and FPGAs. In the tooling category, CUDA has the most specialized and effective tools whereas Vulkan has almost no compute specific tools available. CUDA and Vulkan provide the best performance (on the tested devices), while Vulkan was able to outperform CUDA on a Nvidia Jetson.It becomes clear that all four APIs have their advantages and may be suitable for different applications. Vulkan is the API of choice in mobile consumer applications. CUDA is convenient for fast development but only if it is feasible to support Nvidia GPUs only. For AMD GPU support the CUDA-like HIP can be used, although lacking Windows support and the tool quality on the AMD side. OpenCL might be chosen if a considerable number of different desktop GPUs and FPGAs should be supported. Although some metrics still show room for improvement, SYCL seems to grow into a good all-rounder with all the available implementations. It could benefit from weakening its tight integration with OpenCL to improve support on some platforms. A Vulkan implementation, for example, would open opportunities to target mobile and Apple devices, while hipSYCL could gain the status of a standard-conforming implementation.While these results provide a general and first overview of the differences, the pros, and cons of those APIs, it would be interesting to receive feedback from the Khronos developer community on their experiences and build more mature metrics on it.

Objective: We address the question of necessary conditions for users to adjust system settings, such as alarm thresholds, correctly. Background: When designing systems, we need to decide which system functions users should control. Giving control to users empowers them, but users must have the relevant information and the ability to adjust settings correctly for their control to be beneficial. Method: Using the example of adjusting an alerting threshold, we analyze the conditions for when users can and when they cannot possibly adjust threshold settings adequately. Results: We identify two obstacles that limit users’ ability to adjust thresholds adequately: (a) the difficulty of determining the correct threshold settings, especially because of users’ strong response to false positive indications, and (b) the difficulty of collecting the information necessary for setting the threshold. Conclusion: Users often cannot identify the optimal settings for a system, so it is unlikely that they choose adequate system settings. Application: System designers must consider the difficulties users face and analyze them explicitly when deciding on user involvement in processes.

Users experience an almost continuous stream of information, making it necessary for them to decide to which information to attend and which to ignore. Indications about the relevance of specific information items can help. We report two experiments on the display of the relevance of information in the context of mobile devices, comparing different display formats and levels of resolution of the information. The results show that people prefer relatively few levels of relevance indications (two or three) and that numeric or graphic (bar) formats are usually preferred. We discuss some implications for the design of relevance indicators.

One of the major views of privacy associates privacy with the control over information. This gives rise to the question how controllable privacy actually is. In this paper, we adapt certain formal methods of control theory and investigate the implications of a control theoretic analysis of privacy. We look at how control and feedback mechanisms have been studied in the privacy literature. Relying on the control theoretic framework, we develop a simplistic conceptual control model of privacy, formulate privacy controllability issues and suggest directions for possible research., Comment: The final publication will be available at Springer via http://dx.doi.org/ [in press]

In systems with advanced automation, human responsibility for outcomes becomes equivocal. We developed the Responsibility Quantification (ResQu) model to compute a measure of operator responsibility (Douer & Meyer, 2020) and compared it to observed and subjective levels of responsibility (Douer & Meyer, 2019). We used the model to calculate operators' objective responsibility in a common fault event in the control room in a dairy factory. We compared the results to the subjective assessments made by different functions in the diary. The capabilities of the automation greatly exceeded those of the human, and the operator should comply with the indications of the automation. Thus, the objective causal human responsibility is 0. Outside observers, such as managers, assigned much higher responsibility to the operator, possibly holding operators responsible for adverse outcomes in situations in which they rightly trusted the automation., Submitted version of the paper that was eventually published in the Proceedings of the Human Factors and Ergonomics Society 2020

Dieser Beitrag ist mit Zustimmung des Rechteinhabers aufgrund einer (DFG geförderten) Allianz- bzw. Nationallizenz frei zugänglich. This publication is with permission of the rights owner freely accessible due to an Alliance licence and a national licence (funded by the DFG, German Research Foundation) respectively. Indications from alerts or alarm systems can be the trigger for decisions, or they can elicit further information search. We report an experiment on the tendency to collect additional information after receiving system indications. We varied the proclivity of the alarm system towards false positive or false negative indications and the perceived risk of the situation. Results showed that false alarm-prone systems led to more frequent re-checking following both alarms and non-alarms in the high risk condition, whereas miss-prone systems led to high re-checking rates only for non-alarms, representing an asymmetry effect. Increasing the risk led to more re-checks with all alarm systems, but it had a stronger impact in the false alarm-prone condition. Results regarding the relation of risk and the asymmetry effect of false negative and false positive indications are discussed.

Binary cues help operators perform binary categorization tasks, such as monitoring for system failures. They may also allow them to attend to other tasks they concurrently perform. If the time saved by using cues is allocated to other concurrent tasks, users' overall effort may remain unchanged. In 2 experiments, participants performed a simulated quality control task, together with a tracking task. In half the experimental blocks cues were available, and participants could use them in their decisions about the quality of products (intact or faulty). In Experiment 1, the difficulty of tracking was constant, while in Experiment 2, tracking difficulty differed in the 2 halves of the experiment. In both experiments, participants reported on the NASA Task Load Index that cues improved their performance and reduced their frustration. Consequently, their overall score on mental workload (MWL) was lower with cues. They also reported, however, that cues did not reduce their effort. We conclude that cues and other forms of automation may support task performance and reduce overall MWL, but this will not necessarily mean that users will work less hard. Thus, effort and overall MWL should be evaluated separately, if one wants to obtain a full picture of the effects of automation. (PsycINFO Database Record(c) 2016 APA, all rights reserved). Language: en

Alerting users about possible threats or blocking users’ ability to perform potentially dangerous actions are two common ways to protect systems from the adverse effects of threats, such as malicious email attachments, fraudulent requests, or system malfunctions. We present a normative model of the effects of alerting and blocking on the value of the outcomes, on measures of risk-taking, and on the number of successful attacks. We compared warning and blocking systems and binary- and likelihood-alarm systems as a function of properties of the threats and the security system. We also compared model predictions to actual user behavior, as measured in a controlled experiment. The experimental results were generally in line with the normative model. However, the model predicted that the outcomes from blocking would always be worse or equal to those from warnings. The experiment, however, showed that blocking may have an advantage over warnings, because it leads to fewer undetected events (as predicted by the model), without significantly lowering the mean value of outcomes (the model predicts a lower value). We discuss practical implications regarding the use of blocking and alerting and the more general value of combining optimal decision models and empirical experiments for determining system designs.

Optimality of any decision, including perceptual decisions, depends on the criteria used to evaluate outcomes and on the assumptions about available alternatives and information. In research settings, these are often difficult to define, and therefore, claims about optimality are equivocal. However, optimality is important in applied settings when evaluating, for example, the detection of abnormalities in medical images.

The ‘intuitive’ trust people feel when encountering robots in public spaces is a key determinant of their interactions with the systems. To study the trust we presented subjects with static images of a robot performing an access-control task, interacting with younger and older male and female civilians, applying polite or impolite behavior. Our results showed strong effects of the robot’s behavior. Age and gender of the people interacting with the robot had no significant effect on participants’ impressions of the robot’s attributes. This preliminary study shows that politeness may be a crucial determinant of people’s perception of peacekeeping robots.

Knowledge-based authentication is the oldest and most widely used form of authentication, but it is still problematic. We present a model of the effects of usage cost variables (e.g., code length, required motion precisions) on authentication performance (time for authentication, error rate) and on the decision to use authentication. We tested model predictions in two experiments in which participants played an investment game and had to use authentication to change their investment. We manipulated the authentication method (personal identification number vs. graphical password), the required precision for authentication, the code length, and time pressure. The variables affected authentication decisions and performance, but the effects were not the same. Also, when the graphical password required greater response precision, performance and subjective ratings decreased dramatically, much more than predicted by combining the effects of the variables independently. These results point to a number of issues that must be considered when designing authentication procedures.

There is limited knowledge on how to design effective interfaces to enable nonexpert users to interact with robot learning algorithms. This paper focuses on an interface design challenge: How to provide the user with sufficient information about the learned behavior. A simulated robotic task where the robot has online learning capabilities was developed. This platform was used to study the impact of the variability of the environmental conditions, the information provided on the relation between the learned robot behavior and the conditions in the environment, and the presence of a preview of the learned robot behavior on the use of the system and on task performance. The results show significant effects of the type and the quantity of displayed information. Forty-two participants made the best use of brief and contextualized notifications about changes in the environment: their presence improved the overall performance and the usage of the automation and reduced the workload. In contrast, adding previews of the learned behavior surprisingly impaired performance and reduced the use of automation.

The delivery of personalized news content depends on the ability to predict user interests. We evaluated different methods for acquiring user profiles based on declared and actual interest in various news topics and items. In an experiment, 36 students rated their interest in six news topics and in specific news items and read on 6 days standard, nonpersonalized editions and personalized basic or adaptive news editions. We measured subjective satisfaction with the editions and expressed preferences, along with objective measures, to infer actual interest in items. Users' declared interest in news topics did not strongly predict their actual interest in specific news items. Satisfaction with all news editions was high, but participants preferred the personalized editions. User interest was weakly correlated with reading duration, article length, and reading order. Different measures predicted interest in different news topics. Explicit measures predicted interest in relatively clearly defined topics such as sports, but were less appropriate for broader topics such as science and technology. Our results indicate that explicit and implicit methods should be combined to generate user profiles. We suggest that a personalized newspaper should contain both general information and personalized items, selected based on specific combinations of measures for each of the different news topics. Based on the findings, we present a general model to decide on the personalization of news content to generate personalized editions for readers.

Durations are often used to judge the status of an invisible process. However, the apparent duration of an interval depends on the actual duration and on other variables, such as the workload during the interval and the person's expectations. An experiment dealt with the use of durations as an information source on the state of an invisible process and the effects of expectations and workload on decisions regarding the process. Eighty-nine participants observed a computerized simulation of a process which could be either intact or faulty, with intact processes ending on average sooner than faulty ones, and they had to indicate whether or not the process is intact and to estimate its duration. A binary cue with either intermediate or no validity indicated whether the process was supposedly intact or not, generating expectations about the duration of the process. Perceived durations and the decisions about the intactness of a process depended on the actual process duration, as well as on the expectations generated by the binary cue. In addition, task workload affected time estimates, but it had no effect on participants' tendency to adhere to cue recommendations or their ability to distinguish between intact and faulty processes. Results show that users' duration-based decisions about the status of a computerized process are affected by internal and external cues. While users can use durations as an information source, they should, whenever possible, be accompanied by additional indicators, lowering the inherent uncertainty in the duration estimation process.

Information systems increasingly provide options for visually inspecting data during the process of information discovery and exploration. Little research has dealt so far with user interactions with these systems, and specifically with the effects of characteristics of the displayed data and the user on performance with such systems. The study reports an experiment on users' performance with a visual exploration system. Participants had to identify target graphs within a large set of candidate graphs by using visual filtering criteria that differed in their efficiency for reducing the number of candidate graphs. A pay-off matrix and a time limit served to motivate users to select filter criteria efficiently. Performance was measured as the number of correct identifications of target graphs within a time-limit, and the number, type and position of filter criteria selected for the search. Efficiency was somewhat biased by users' preference to select filter criteria sequentially, starting from the left to t...

The editor's decision where and how to place items on a screen is crucial for the design of information displays, such as websites. We developed a statistical model that can facilitate automating this process by predicting the perceived importance of screen items from their location and size. The model was developed based on a 2-step experiment in which we asked participants to rate the importance of text articles that differed in size, screen location, and title size. Articles were either presented for 0.5 seconds or for unlimited time. In a stepwise regression analysis, the model's variables accounted for 65% of the variance in the importance ratings. In a validation study, the model predicted 85% of the variance of the mean apparent importance of screen items. The model also predicted individual raters' importance perception ratings. We discuss the implications of such a model in the context of automating layout generation. An automated system for layout generation can optimize data presentation to suit users' individual information and display preferences. © 2011 Wiley Periodicals, Inc.

In this paper, we describe a new approach to analyze the trade-off between usability and security frequently found in security-related user interfaces. The approach involves the simulation of potential user interaction behavior by a mixed probabilistic and rule-driven state machine. On the basis of the simulations, user behavior in security-relevant situations can be predicted and user interfaces optimizing intended behavior can be designed. The approach is evaluated in an artificial microworld setting which provides good control over the experimental factors guiding the behavior. A comparison of empirical and simulated behavior in this microworld shows that the approach is already able to accurately predict important aspects of user behavior toward security interfaces, but also identifies future work necessary to better cover all relevant aspects guiding this behavior in a real-world setting.

Personalizing news content requires to choose the appropriate depth of personalization and to assess the extent to which readers' explicit expressions of interest in general and specific news topics can be used as the basis for personalization. A preliminary survey examined 117 respondents' attitudes towards news content personalization and their interest in various news topics and subtopics. The second survey examined 23 participants' declared and actual interests. Participants preferred personalization based on general news topics. Declared interest in general news topics adequately predicted the actual interests in some topics, while in others users' interests differed between general news topics and subtopics. The variance in interest in items also differed among topics. Thus, different personalization methods should be used for different topics. For some, such as 'Sports', users show either high interest or no interest at all. In the latter case most articles related to the topic should be removed, with the exception of items that refer to unique events that may raise general interest according to the expressed interest. In other topics, such as 'Science & Technology', most users are interested in important articles, even if they are not interested in the general news topic. Here, the filtering technique should identify the important articles and present them to all readers. The results can be used to develop effective and simple personalization mechanisms which can be applied to the personalization of news, as well as to other domains.

The paper examines the positive and the possible adverse effects of adaptive user interfaces (AUIs) in the context of an in-vehicle telematic system as a function of four factors: (1) four different levels of adaptivity (ranging from manual to fully adaptive with intermediate levels); (2) different tasks; (3) routine (familiar) and non-routine (unfamiliar) situations; and (4) different user age groups. Both experiments included three sessions during which participants drove a simple driving simulator and performed tasks with the telematic system at one of the adaptivity levels. We measured task performance times and lane position variance. Adaptivity was not always equally beneficial, and its benefits depended on a number of factors, including the frequency in which the tasks were performed, the user's age, the difficulty of the task and the user's involvement in the task. In familiar, routine situations, a fully adaptive system was beneficial for all participants, particularly older ones. In unfamiliar situations, to which the AUI was not adjusted, cognitive workload increased substantially, adversely affecting performance. Intermediate levels of adaptivity keep users involved in the task and help them become more proficient when performing both routine and non-routine tasks. However, intermediate levels of adaptivity should also be implemented with care, because they may also have adverse effects when users encounter non-routine situations.

Robotic systems in unstructured environments must cope with unknown, unpredictable, and dynamic situations. Inherent uncertainty, and limited sensor accuracy and reliability impede target recognition performance. Introducing a human operator into the system can help improve performance and simplify the robotic system. In this paper, four basic levels of collaboration were defined for human-robot collaboration in target recognition tasks. An objective function that includes operational and time costs was developed to quantify performance and determine the best collaboration level. Signal detection theory was applied to evaluate system performance. The optimal collaboration level for different cases was determined by using numerical analyses of the objective function. The findings indicate that the best system performance, the optimal values of performance measures, and the best collaboration level depend on the task, the environment, human and robot parameters, and the system characteristics. For the tested cases, the manual level was never the best collaboration level for achieving the optimal solution. The autonomous level was the best collaboration level when robot sensitivity was higher than human sensitivity. In general, collaboration of human and robot in target recognition tasks will improve upon the optimal performance of a single human detector.

In two experiments participants had to detect changes in periodic sinusoidal functions, displayed in either graphic or tabular displays. Graphs had a major advantage over tables when the task required considering configurations of data. Both displays led to similar results when task performance could rely on inspecting individual data points. With graphs almost all participants reported using the optimal method for detecting changes in the function, i.e., they used the method requiring the least effort to perform the task. With tables only about half used the optimal detection method, and these participants showed transfer of learning of detection methods between tasks. Experience in using a detection method led to improved performance if the new task relied on the same method of detection. These findings demonstrate the need to consider task performance methods when determining the relative value of different displays. The set of tasks for which a display is used is likely to affect performance and needs to be analysed as a whole, since methods employed for one task can affect task performance in other tasks.

People make numerous decisions that affect their own or others’ privacy, including the decisions to engage in certain activities, to reveal and share information or to allow access to information. These decisions depend on properties of the information to be revealed, the situation in which the decision is made, the possible recipients of the information, and characteristics of the individual person. System design should ideally protect users from unwanted consequences by allowing them to make informed decisions, at times blocking users’ ability to perform certain actions (e.g., when the user is a minor). The development of alerting and blocking mechanisms should be based on predictive models of user behavior, similar to engineering models in other domains. These models can be used to evaluate different design alternatives and to assess the required system specifications. Predictive models of privacy decisions will have to combine elements from normative decision making and from behavioral, descriptive research on decision making. Some major issues in the development and validation of such models are presented.

Cues in visual scanning task can improve decision accuracy, and they may also affect task performance strategies. We tested the effects of cues on the performance of binary classifications, on the screen scanning procedure participants employed, and on the reported effort in a simulated quality control task. Participants had to decide whether each item in a 5 × 5 matrix of items was intact or faulty. In half the experimental blocks decisions could only be based on the visual properties of the items. In the other half, participants also saw imperfect binary cues and could use them to classify the items as faulty or intact. We used eye tracking to study scan patterns and fixation durations on items. Decision performance improved with cues, and cues affected the scanning of items, with participants mainly scanning cued items and tending to scan them longer. Participants stated that cues reduced their effort when cues were highly valid. We conclude that strategic choices to focus on suspected areas determined the screen scanning procedure, the amount of effort invested in single decisions, and the accuracy of the decisions. We therefore suggest using likelihood ratio cues to help optimize the scanning procedure.

Decisions in almost all domains of life receive support from automation in the form of alerts, binary cues, recommendations, etc. People often use automation or decision aids without having experience with the system, because the system may be new or because they rarely use it. When such experience is unavailable, people will base their use of the system on information they may have received about it and on descriptions, often given as probabilities or proportions. Examples are the sensitivity and specificity of a diagnostic procedure in medicine or the True Positive and False Positive rates of a detector. People use these descriptions to decide to what extent they can rely on the information. So far, it is unclear which aspects of the information about a system determine people’s evaluation of the system from a description. These evaluations will determine the trust they put in the indications from the system and the adjustment of system properties, such as thresholds. To gain some insights into this issue, we conducted an experiment. We developed descriptions of 12 systems in a quality control setting, in which participants had to detect faulty items in a production process. We used Signal Detection Theory (Green & Swets, 1966) to determine the system properties. The systems differed in d’ (1.5 or 2.5), the threshold setting lnβ (-1, 0 or 1) and the prior probability for a signal pS (.05 or .2). Half of the participants saw diagnostic values, receiving descriptions in terms of the probabilities of Hit and False Alarms, while the other half saw descriptions as predictive values, receiving the Positive Predictive Value (PPV) and the Negative Predictive Value (NPV) of each system. In the past, we have shown that people adjust system thresholds better when they see predictive values (Botzer, Meyer, Bak, & Parmet, 2010). Fifty-six students evaluated the systems in a classroom setting on a scale between 0 (completely useless) and 10 (perfect). In addition to the d’ and lnβ, which we specified when we designed the systems, we also computed for each system the Probability of Correct Indication (pCorrect), the Expected Value (given the costs and benefits in the description), and the transmitted information according to Information Theory. We analyzed the results with multivariate analyses of variance and by computing the correlations between the evaluations and system properties. The results showed that participants’ responses were mainly correlated with d’. The effects of the threshold setting lnβ and of pS were small, compared to the effects of d’. The correlations with the Expected Value and the transmitted information were smaller and could be explained through d’. Thus, people evaluated a system in terms of its ability to differentiate between signal and noise. They did not evaluate the system according to the economic value it provided or the transmitted information. In addition, participants evaluated systems with different thresholds (lnβ) similarly. This means that in our experiment participants did not differentiate between more and less appropriate threshold settings. The ability to identify better or worse settings is important, because these settings are often the main system parameter users can adjust. These findings, in addition to the inherent problems that already exist in user adjustments of systems (Meyer & Sheridan, 2017), make it unlikely that people can adjust system settings correctly.

Three experiments assessed the relative efficiency of graphic and tabular displays for detecting changes in periodic sine functions, simulating a dynamic process. In all three experiments, line graphs had an advantage over tables for response times, for correct detections and identifications, and for false alarms. However, this advantage depended on the type of change that indicated a malfunction. The largest difference between the displays was evident for the detection of changes in a function's frequency and smaller differences were found for the detection of changes in amplitude and intercept. Results also indicated that participants adapted their detection methods to the types of possible changes. The findings demonstrate the value of graphic displays for process control and substantiate the claim that graphic displays have an advantage when the displayed information has inherent structure and when the task requires the use of this structure. In addition, task performance was the subject of adaptive c...

The relative efficiency of graphic and tabular displays for detecting changes in the amplitude of periodic sine functions, simulating a dynamic process, was assessed. Line graphs had an advantage over tables for the detection of changes and for the correct identification of the changed function. However, the advantage depended on the type of change that could occur. A large difference between the displays was evident when both increases and decreases in amplitude were possible, and differences were much smaller when amplitudes could only increase. These results indicate that participants adapted their detection methods to the types of possible changes. The findings demonstrate the utility of graphic displays for process control and substantiate the claim that graphic displays have an advantage when the displayed information has inherent structure and when the task requires the use of this structure. In addition, task performance was shown to be the subject of adaptive changes, which depend on the context in which the task is performed.

Endoscopic surgery, while offering considerable gains for the patient, has created new difficulties for the surgeon. One problem is the fulcrum effect, which causes the movement of a surgical instrument, as seen on the monitor, to be in the opposite direction to the movement of the surgeon's hand. The problem has been shown to impede the acquisition of endoscopic skills. Teleoperated robotic arms may circumvent this problem by allowing different control-response relations. Four alternative control designs of a teleoperated device were compared in a simulated endoscopic task. A rigid teleoperated robotic arm with two degrees of freedom representing a surgical tool was coupled to a joystick in a position control mode. Feedback was provided through a video display. Participants without prior experience in endoscopy performed a target acquisition task, first by pointing the robotic arm at the targets, and later by maneuvering an object. Performance was measured under four different combinations of visual-motor mapping (normal/reversed), and the joystick's orientation (upwards/downwards). Task completion time under normal visual-motor mapping was found to be significantly shorter than under reversed visual-motor mapping, emphasizing the potential advantage of a teleoperated endoscopic system. The joystick's orientation affected the maneuvering of an object under only the reversed visual-motor mapping, implying that the positioning of a surgical tool and the manipulation of tissues or objects with the tool may be differentially affected by the control design.

Users and system designers often prefer to display information with graphs rather than with tables. However, empirical studies that compared task performance with the two display types frequently revealed either an advantage of tables over graphs or no differences between the displays. This apparent contradiction may result from previous studies in which the importance of the structure that usually exists in displayed information is overlooked. We predict that graphic displays will have an advantage over tables when the displayed information has structure and when this structure is relevant for the task. These conditions generally exist in the actual use of information displays, but have seldom been assessed in experiments. In the present study participants in an experiment performed an information extraction task and a prediction task with unstructured or structured data and with different levels of prior information about the structure. The results showed that the information structure and prior knowledge about the existence of structure affected the advantage of graphic displays over tables when task performance depended on the use of structure. Existing approaches to the study of displays were analyzed in view of these findings. Actual or potential applications of this research include the development of better displays for process control and decision support and better operator training programs.

An experiment assessed the relative efficiency of tables and graphs for the discrimination between two system states, indicated by different frequencies of sine functions, in a simulated process. Graphic displays allowed significantly better discrimination than tables. The interval between the feed back for one trial and the beginning of the next trial affected performance for both displays. In graphs rapid learning occurred for all durations of the interval, but when the interval was 0 seconds the initial performance was lower than when the interval was 2 or 6 seconds. For tables only the 6 seconds interval showed faster learning than the other conditions, but performance was still much lower than for graphs. The results lend support to the claim that graphic displays are of advantage when the displayed data has structure and when the task relies on this structure.

Washburne (1927) conducted the largest study to date on the display of quantitative information with tables and graphs. Measuring the accuracy of information recall, he concluded that bar graphs are particularly suited for complex comparisons between values, pictographs are best for simple comparisons, line graphs are best for identifying and comparing trends, and tables are best for specific values. These conclusions have intuitive appeal and were quoted in the literature on comparative displays. A reanalysis of the results indicates that only the claim that tables are best for recalling specific values is clearly supported by Washburne's data. The other conclusions were probably accepted without questioning because they appeal to intuition. Additional findings of the study, such as the importance of the logical ordering of items in a display, were not acknowledged in subsequent studies. In human factors, as in other fields, naive preconceptions exist which can impede the acquisition of new knowledge.

Two experiments assessed the relative efficiency of line graphs, bar graphs, and tables, applying a multiple-factors approach to study the effects of the type of the required information, the complexity of the data, and the user's familiarity with the display. information extraction tasks included reading exact values, comparing values, identifying trends, and reading maximum values. Tables led to faster responses for all tasks, and the accuracy for tables was equally high or higher than for graphs. Bar graphs and line graphs differed in their relative efficiency for the different tasks. The complexity of the data also affected the tasks differentially, as did prior familiarity with the display. Performance for most conditions improved with experience. Our findings demonstrate the benefits of a multiple-factors approach to the study of displays. Generalizations about the relative efficiency of displays and computational models of the task performance with displays must consider the various relevant factors if they are to serve as valid design aids.

Binary cueing systems assist in many tasks, often alerting people about potential hazards (such as alarms and alerts). We investigate whether cues, besides possibly improving decision accuracy, also affect the effort users invest in tasks and whether the required effort in tasks affects the responses to cues. We developed a novel experimental tool to study binary categorization performance. In two experiments, participants decided whether items on a screen were intact or faulty, based on the configuration of lighter and darker areas in items. Cues were available in half of the experimental blocks, and participants could use them in their decisions. Experimental conditions differed in the effort required to perform the task, manipulated through the contrast between lighter and darker areas (higher contrast vs. lower contrast), and in the validity of cues (medium vs. high validity). In the NASA-TLX, participants reported that with highly valid cues, they invested less effort in the task, whereas with medium validity cues, they invested similar effort as without cues. Responses to the high-validity cues were stronger than responses to the medium-validity cues. The required effort in the task did not affect the strength of responses to cues. We conclude that the invested effort may decrease when cues are available, but users will not rely more strongly on cues in more demanding situations to reduce the invested effort. We therefore recommend integrating cues into work environments, as they reduce users' effort without necessarily leading to overdependence on the cues to reduce invested effort.

The issue of optimal display design was re-examined under the hypothesis that the nature of the data set would influence the efficacy of displays. An experiment assessed the effect of data structure on the relative efficacy of line graphs and tables. Participants saw a sequence of graphic or tabular displays which presented data from functions. The displays showed either values taken from the simple form of a single sinusoid function (structured data conditions), or from the complex and seemingly unstructured form of five summed sinusoid functions (unstructured data conditions). After seeing four consecutive displays participants were asked to predict the behavior of the data set in the next display which they had not yet seen. They were required to predict either the behavior of specific points, or the direction of trends. While the performance of the point comparison task was not influenced by display format, graphs were found to have an advantage for the prediction of future trends. Graphs also led more frequently to the identification of structure than did tables, both for structured and unstructured conditions.

The way people intuitively manage inventories is both an important issue in its own right and a context for research on various basic theoretical questions. The subject has not yet been systematically studied, owing partially to the lack of appropriate research tools. This paper presents some of the key concepts in formal inventory theory and describes an experimental system for the study of intuitive inventory management. This system is a computer program that allows subjects to interact with a simulated inventory system. Numerous system parameters can be manipulated, including demand functions, prices, costs, and the display of data. The usefulness of the system is demonstrated via the study of the effects of three parameters on intuitive inventory management in an illustrative experiment. Possible future uses of the system in basic and applied research and in training are indicated.

Two experiments assessed the effect of displaying a boundary on duration estimates and preference ratings for dynamic displays that were shown while users waited for the system's response. Displays were either symbolic (changing numbers) or graphic (increasing rectangles) and could contain a boundary that indicated when the interval was expected to be over. Duration estimates were similar for symbolic and graphic displays and for displays with and without a boundary. However, when the displays were encountered successively, participants assessed the graphic displays as having shorter durations than the symbolic displays. Faster rates of change in both types of displays led to increased duration estimates. Although displaying a boundary did not affect duration estimates, participants preferred displays in which a boundary was shown and preferred the graphic displays over the symbolic displays. Hence, bounded graphic displays are recommended as “wait” displays for computerized applications.

Mobile phones are rapidly becoming small-size general purpose computers, so-called smartphones. However, applications and data stored on mobile phones are less protected from unauthorized access than on most desktop and mobile computers. This paper presents a survey on users' security needs, awareness and concerns in the context of mobile phones. It also evaluates acceptance and perceived protection of existing and novel authentication methods. The responses from 465 participants reveal that users are interested in increased security and data protection. The current protection by using PIN (Personal Identification Number) is perceived as neither adequate nor convenient in all cases. The sensitivity of data stored on the devices varies depending on the data type and the context of use, asking for the need for another level of protection. According to these findings, a two-level security model for mobile phones is proposed. The model provides differential data and service protection by utilizing existing capabilities of a mobile phone for authenticating users.

One way to secure Information Technology (IT) systems is with authentication mechanisms that distinguish between users. Users who differ in their cognitive and motor abilities, cultural background and personal characteristics should be able to operate the IT system including its security features. If system design fails to consider user diversity, users might bypass or disable the security feature, reducing system security. Providing universal accessesibility and acceptability is generally a challenge, especially when dealing with IT security. We present a conceptual model that explores and establishes guidelines for the inclusion of biometric authentication in systems which serve a wide range of users. Aspects of this model were examined in laboratory settings using a task which simulates mobile access to an eBanking system with biometric authentication. Younger and older participants used the authentication mechanism. The age groups clearly differed in their interaction with the IT and the security system. Designing security system for universal access remains a major challenge.

With robots implementing learning capabilities, a whole new range of issues regarding the human interaction with the learning process arises. Among these issues, the question of the nature of the interaction while the human is teaching the robot is critical. To solve the problem of interface design for teaching robots, we propose a new approach based on Gibson’s theory of visual perception [4].

The car is rapidly changing. In addition to its traditional driving-related functions it becomes a platform for various services and devices. Some of these are related to the driving task and can improve its ease, comfort and safety, while others are unrelated to driving, and allow the driver to engage in various activities while driving. The aging of the driving population, and the tendency of older people in many parts of the world to continue driving for as long as possible, pose major challenges regarding the design of such devices and their deployment in cars. Some advantages, as well as some limitations which these devices may have for older drivers are pointed out. Design of future in-vehicle telematic systems will have to consider these issues in order to provide maximum benefits for the older driver.

The aim of this study was to examine how users perceive help on a mobile device with respect to the presentation format and the severity of the scenario the user encounters. We examined how 92 participants responded to four different formats of help offered for five different scenarios varying in their severity level using static images displayed in a web browser. The intrusiveness of the help format and the severity of the scenario affected participants' interest in receiving help. In certain conditions it may be helpful to suggest help, in particular when users do not know how to ask for it but show interest in receiving it when offered.

Dieser Beitrag ist mit Zustimmung des Rechteinhabers aufgrund einer (DFG geförderten) Allianz- bzw. Nationallizenz frei zugänglich. This publication is with permission of the rights owner freely accessible due to an Alliance licence and a national licence (funded by the DFG, German Research Foundation) respectively. The purpose of this panel was to provide a general overview and discussion of some of the most current and controversial concepts and trends in human-automation interaction. The panel was composed of eight researchers and practitioners. The panelists are well-known experts in the area and offered differing views on a variety of different human-automation topics. The range of concepts and trends discussed in this panel include: general taxonomies regarding stages and levels of automation and function allocation, individualized adaptive automation, automation-induced complacency, economic rationality and the use of automation, the potential utility of false alarms, the influence of different types of false alarms on trust and reliance, and a system-wide theory of trust in multiple automated aids.

An ideal system should be usable and secure. However, increasing the security of a system often makes its use more cumbersome and less efficient. This tradeoff between usability and security poses major challenges for system designers. System security may be impaired when users override or ignore security features to facilitate the use of the system. Little empirical data are available on user behavior regarding the tradeoff between security and usability. To obtain such data we developed a controlled research environment (i.e., a microworld) for studying users’ tendency to take precautionary actions as a function of the tradeoff between a system’s usability and the level of security the system provides. It is a modified version of a “Tetris” game and includes an alert system that warns about possible virus attacks, which, if not prevented, can cause losses of monetary earnings. Users could alter the threshold settings of the security system. The system allows us to manipulate the usability cost of using a security feature, the severity of the consequences of an attack, the likelihood that a threat will occur, and the statistical properties of the security system. In a preliminary experiment two groups of 10 participants each used the system for three 20-minutes sessions. The likelihood for an attack was 4 times higher for one group than for the other group. The likelihood of an attack clearly affected participants' behavior. When attacks were more likely, participants altered thresholds more frequently, selected more cautious thresholds, and tended to respond more to security system alerts. This microworld is a step towards the development of quantitative predictive models of user interactions with security features while using a system.

We describe a design project of a future electronic news device and service. The project employs about 20 researchers, designers and developers. It uses advances in product technologies and in social computing to deal with the challenges of transferring the print newspaper reading experience onto a mobile, hand-held device, and of transferring the editor's power to decide what constitutes worthy news to the reader.

Motivation -- To address the question how the effort required to obtain task relevant information and the sensitivity of a binary cueing system affect responses to cues in a binary categorization task that is aided by a binary cueing system.Research approach -- In a simulated quality control task, participants had to decide whether to discard an item or approve it based on the configuration of light and darker squares in it. In half the experimental blocks they could use cues from a binary cueing system. The experimental conditions differed in the contrast between the light and darker squares in the items and in the sensitivity of the cueing system. At the end of each experimental block participants completed the NASA-TLX questionnaire to assess their perceived workload.Findings -- The results showed that lower contrast led to higher perceived effort. However, compliance with the cues and reliance on them were not stronger among participants for whom the task required more effort. In addition, the use of binary cues reduced perceived workload only with the more sensitive binary cueing system.Take away message -- The results may suggest that when not overloaded people may prefer to be involved in decisions instead of letting them be strongly guided by an automated device.