Human-machine Teaming Research Articles

Human Performance in Competitive and Collaborative Human-Machine Teams.

In the modern world, many important tasks have become too complex for a single unaided individual to manage. Teams conduct some safety-critical tasks to improve task performance and minimize the risk of error. These teams have traditionally consisted of human operators, yet, nowadays, artificial intelligence and machine systems are incorporated into team environments to improve performance and capacity. We used a computerized task modeled after a classic arcade game to investigate the performance of human-machine and human-human teams. We manipulated the group conditions between team members; sometimes, they were instructed to collaborate, compete, or work separately. We evaluated players' performance in the main task (gameplay) and, in post hoc analyses, participant behavioral patterns to inform group strategies. We compared game performance between team types (human-human vs. human-machine) and group conditions (competitive, collaborative, independent). Adapting workload capacity analysis to human-machine teams, we found performance under both team types and all group conditions suffered a performance efficiency cost. However, we observed a reduced cost in collaborative over competitive teams within human-human pairings, but this effect was diminished when playing with a machine partner. The implications of workload capacity analysis as a powerful tool for human-machine team performance measurement arediscussed.

Estimating Systemic Cognitive States from a Mixture of Physiological and Brain Signals.

As human-machine teams are being considered for a variety of mixed-initiative tasks, detecting and being responsive to human cognitive states, in particular systematic cognitive states, is among the most critical capabilities for artificial systems to ensure smooth interactions with humans and high overall team performance. Various human physiological parameters, such as heart rate, respiration rate, blood pressure, and skin conductance, as well as brain activity inferred from functional near-infrared spectroscopy or electroencephalogram, have been linked to different systemic cognitive states, such as workload, distraction, or mind-wandering among others. Whether these multimodal signals are indeed sufficient to isolate such cognitive states across individuals performing tasks or whether additional contextual information (e.g., about the task state or the task environment) is required for making appropriate inferences remains an important openproblem. In this paper, we introduce an experimental and machine learning framework for investigating these questions and focus specifically on using physiological and neurophysiological measurements to learn classifiers associated with systemic cognitive states like cognitive load, distraction, sense of urgency, mind wandering, and interference. Specifically, we describe a multitasking interactive experimental setting used to obtain a comprehensive multimodal data set which provided the foundation for a first evaluation of various standard state-of-the-art machine learning techniques with respect to their effectiveness in inferring systemic cognitive states. While the classification success of these standard methods based on just the physiological and neurophysiological signals across subjects was modest, which is to be expected given the complexity of the classification problem and the possibility that higher accuracy rates might not in general be achievable, the results nevertheless can serve as a baseline for evaluating future efforts to improve classification, especially methods that take contextual aspects such as task and environmental states into account.

Shaped-Charge Learning Architecture for the Human-Machine Teams.

In spite of great progress in recent years, deep learning (DNN) and transformers have strong limitations for supporting human-machine teams due to a lack of explainability, information on what exactly was generalized, and machinery to be integrated with various reasoning techniques, and weak defense against possible adversarial attacks of opponent team members. Due to these shortcomings, stand-alone DNNs have limited support for human-machine teams. We propose a Meta-learning/DNN → kNN architecture that overcomes these limitations by integrating deep learning with explainable nearest neighbor learning (kNN) to form the object level, having a deductive reasoning-based meta-level control learning process, and performing validation and correction of predictions in a way that is more interpretable by peer team members. We address our proposal from structural and maximum entropy production perspectives.

PART I: The Singleton Paradox: On the Future of Human-Machine Teaming and Potential Disruption of War Itself

Technological innovation has historically been applied in war and security affairs as a new tool or means to accomplish clear political or societal goals. The rise of artificial intelligence posits a new, uncharted way forward that may be entirely unlike previous arms races and advancements in warfare, including nuclear weapons and quantum technology. This article introduces the concept of a singleton as a future artificial intelligent entity that could assume central decision making for entire organizations and even societies. In turn, this presents what is termed a “singleton paradox” for security affairs, foreign policy, and military organizations. An AI singleton could usher in a revolutionary new world free of war and conflict for all of human civilization or trigger a catastrophic new war between those with a functioning singleton entity against those attempting to develop one, along with myriad other risks, opportunities, and emergent consequences.

Development of a machine learning-based method for the analysis of microplastics in environmental samples using µ-Raman spectroscopy

This research project investigates the potential of machine learning for the analysis of microplastic Raman spectra in environmental samples. Based on a data set of > 64,000 Raman spectra (10.7% polymer spectra) from 47 environmental or waste water samples, two methods of deep learning (one single model and one model per class) with the Rectified Linear Unit function (ReLU) (hidden layer) as the activation function and the sigmoid function as the output layer were evaluated and compared to human-only annotation. Based on the one-model-per-class algorithm, an approach for human–machine teaming was developed. This method makes it possible to analyze microplastic (polyethylene, polypropylene, polystyrene, polyvinyl chloride, and polyethylene terephthalate) spectra with high recall (≥ 99.4%) and precision (≥ 97.1%). Compared to human-only spectra annotation, the human–machine teaming reduces the researchers’ time required per sample from several hours to less than one hour.

Super Mario Meets AI: Experimental Effects of Automation and Skills on Team Performance and Coordination

Abstract This article studies the effects of the introduction of artificial intelligence into teams in a laboratory experiment. We randomly assign automated co-workers into “laboratory firms” who are playing a team-based game on the Nintendo Switch console. We demonstrate that even in a task where AI outperforms humans, automation decreases overall team performance and increases coordination failures. These effects are especially large in the short-term and in low- and medium-skilled teams. Moreover, automation reduces team trust and individual effort provision. Our results support the implication that improving collaborative human-machine teams is key to the positive transformation that AI may bring to organizations.

When will AI misclassify? Intuiting failures on natural images.

Machine recognition systems now rival humans in their ability to classify natural images. However, their success is accompanied by a striking failure: a tendency to commit bizarre misclassifications on inputs specifically selected to fool them. What do ordinary people know about the nature and prevalence of such classification errors? Here, five experiments exploit the recent discovery of "natural adversarial examples" to ask whether naive observers can predict when and how machines will misclassify natural images. Whereas classical adversarial examples are inputs that have been minimally perturbed to induce misclassifications, natural adversarial examples are simply unmodified natural photographs that consistently fool a wide variety of machine recognition systems. For example, a bird casting a shadow might be misclassified as a sundial, or a beach umbrella made of straw might be misclassified as a broom. In Experiment 1, subjects accurately predicted which natural images machines would misclassify and which they would not. Experiments2 through 4 extended this ability to how the images would be misclassified, showing that anticipating machine misclassifications goes beyond merely identifying an image as nonprototypical. Finally, Experiment 5 replicated these findings under more ecologically valid conditions, demonstrating that subjects can anticipate misclassifications not only under two-alternative forced-choice conditions (as in Experiments 1-4), but also when the images appear one at a time in a continuous stream-a skill that may be of value to human-machine teams. We suggest that ordinary people can intuit how easy or hard a natural image is to classify, and we discuss the implications of these results for practical and theoretical issues at the interface of biological and artificial vision.

A hybrid cognitive model for machine agents

Collaborative teams pursue common goals, completing tasks and making decisions with various levels of interdependence. A shared mental model (SMM) is a foundational structure in high performing, production teams and aids humans in predicting their teammate’s goals and intentions. Advice teams utilize a transactive memory system (TMS) that integrate sources of knowledge and the source’s credibility. SMMs and TMSs elevate human performance when the nature of emergence complements the associated team type. However, project and action teams require both behavioral and knowledge integration. We present a hybrid cognitive model (HCM) for machine agents that unifies SMM and TMS characteristics. The HCM enables anytime selection over the two cognitive representations with the computational complexity of a single model. Furthermore, the pliant nature of credibility modeling in TMSs can represent expertise, thoroughness, or trust simultaneously in the team context. Results in a multi-agent project domain demonstrate the HCM’s efficacy for machine agent teams and potential for applications in human–machine teams.

Editorial: Interdisciplinary approaches to the structure and performance of interdependent autonomous human machine teams and systems

Editorial: Interdisciplinary approaches to the structure and performance of interdependent autonomous human machine teams and systems

The Inner Loop of Collective Human-Machine Intelligence.

With the rise of artificial intelligence (AI) and the desire to ensure that such machines work well with humans, it is essential for AI systems to actively model their human teammates, a capability referred to as Machine Theory of Mind (MToM). In this paper, we introduce the inner loop of human-machine teaming expressed as communication with MToM capability. We present three different approaches to MToM: (1) constructing models of human inference with well-validated psychological theories and empirical measurements; (2) modeling human as a copy of the AI; and (3) incorporating well-documented domain knowledge about human behavior into the above two approaches. We offer a formal language for machine communication and MToM, where each term has a clear mechanistic interpretation. We exemplify the overarching formalism and the specific approaches in two concrete example scenarios. Related work that demonstrates these approaches is highlighted along the way. The formalism, examples, and empirical support provide a holistic picture of the inner loop of human-machine teaming as a foundational building block of collective human-machine intelligence.

A connective framework to minimize the anxiety of collaborative Cyber-Physical System

ABSTRACT The role of Cyber-Physical systems (CPS) is well recognized in the context of Industry 4.0, which consists of human operators working with machines/robots. The interactions among them can be quite demanding in terms of cognitive resources. Existing systems do not yet consider the psychological aspects of safety in the domain. This lack can lead to hazardous situations, thus compromising the performance of the working system. This work proposes a connective decision-making framework for a flexible CPS, which can quickly respond to dynamic changes and be resilient to emergent hazards. First, Anxiety is defined and categorized for expected/unforeseen situations that a CPS could encounter through historical data using the Ishikawa method. Second, visual cues are used to gather the CPS’s current state (such as human pose and object identification). Third, a mathematical model is developed using Mixed-integer programming (MIP) to allocate optimal resources, to tackle high-impact situations generating Anxiety. Finally, the logic is designed for an effective counter-mechanism to mitigate Anxiety. The proposed method was tested on a realistic industrial scenario incorporating a collaborative CPS. The results demonstrated that the proposed method improves the decision-making of a CPS facing a complex scenario, ensures physical safety, and effectively enhances the human-machine team’s productivity.

Conversations in Human-Machine Teams - Are We On the Record?

Most of us have blithely accepted the terms and conditions of software products we use in our private and business life. This generally hasn’t been that important as we assume a mild erosion of our privacy in exchange for access to services we want or need. However, when translating the issue to human-machine teams for earning our livelihood we will need to pay much more attention to what we are agreeing to. When the future pilot signs for their aircraft and agree to the terms and conditions for using a futuristic AI-based aircraft, they had better understand what they are really signing for – their job may depend on it.

The Future Combat Pilot - Asleep at the Wheel?

Technology is a key driver for Air Force capability. There is always debate on the future technology that can advance our future combat aircraft such as artificial intelligence (AI), quantum computing and stealth technology. But what about the human in this human-machine team? There appears to be little debate about the humans in this team.

ЧИ МОЖЕ ЗАМІНИТИ НЕЙРОМЕРЕЖА CHATGPT РОБОТУ МАРКЕТОЛОГА?

In the process of transforming the Internet from a global communication tool into an effective tool in the field of marketing, the industry where everything changes almost every day, and therefore requires a marketer to familiarize himself with marketing channels and chats, read online publications and develop new hard skills. This was the challenge for the fact that after GPT chat entered the market, it was almost entirely associated with the work of a marketer. Of course, it has undeniable potential: Bot can create content, generate ideas, and take over routine work. According to a survey in the US, 37% of marketing and advertising professionals said they have already used ChatGPT and other AI-based tools in their work. So, a bot can bring a lot of benefit specifically to content marketing: improving the text, creating interesting and relevant content to attract attention and attract a new audience, adapting it to the target audience, communicating with customers through round-the-clock support, automating routine tasks. In our opinion, ChatGPT is designed to help humans by performing simple and routine work, while the human beings can devote themselves to something more important and creative. The human control over the neural network proves the importance of the human factor, as the text generated by ChatGPT may not correspond to the message or interests that the company wants to convey to consumers, its data is limited to 2021, and it has no self-awareness. After all, ChatGPT is not an expert in a particular field, it can only make recommendations, and the final decision is always up to a professional. The development of artificial intelligence is causing changes in the labor market, with professions responsible for routine and monotonous duties and dangerous human labor. However, the emergence of new professions, such as: digital removalist, data detective, drone manager, chief automation officer, rewilder, human-machine teaming manager, Internet lawyer digital currency advisor, smart home design manager, and trash engineer/garbage designer all suggest that artificial intelligence will not take away jobs from people, but rather offer to replace them with more progressive areas.

Battlefield Trust for Human-Machine Teaming: Evidence from the US Military

Battlefield Trust for Human-Machine Teaming: Evidence from the US Military

Examining autonomous flight safety systems from a cognitive systems engineering perspective: Challenges, themes, and outlying risks

Flight safety systems (FSS) act as a method to terminate off-nominal rocket launches which threaten public safety. Traditional FSS delegate decision authority to an experienced Missile Flight Control Officer (MFCO) tasked with flight termination decisions, observing multiple points of telemetry data in real time to ensure nominal flight status. This study examines the engineering trade-offs, complexities, and pitfalls introduced by automating flight termination decision-making through autonomous flight safety systems (AFSS). We approach this problem from a cognitive systems engineering perspective, connecting aspects of AFSS to existing literature in human-machine teaming and resilience engineering. Based on information gathered from a series of semi-structured interviews with various subject matter experts (mission controllers, regulators, engineers, amongst others) and existing literature, we outline a list of four assumptions underlying AFSS operations: (1) The system is fully autonomous, (2) An exhaustive flight safety analysis has been performed, (3) The system will be able to respond appropriately to the world, and (4) MFCO expertise can be captured in (or translated to) software. Our findings highlight that, while the benefits of AFSS hold great promise for increasing the viability of commercial space operations, the automation of an irreversible, instantaneous, and complex decision-making task introduces significant challenges and risks. We propose directions for further research to minimize the likelihood of errant, expensive, and dangerous automated flight terminations.

Interdependence design principles in practice

Adaptability lies at the heart of effective teams and it is through management of interdependence that teams are able to adapt. This makes interdependence a critical factor of human-machine teams. Nevertheless, engineers building human-machine systems still rely on the same tools and techniques used to build individual behaviors which were never designed to address the complexity that stems from interdependence in joint activity. Many engineering approaches lack any systematic rigor and formal method for identifying, managing and exploiting interdependence, which forces ad hoc solutions or workarounds. This gap between theories of interdependence and operable tooling leaves designers blind to the issues and consequences of failing to adequately address interdependence within human-machine teams. In this article, we propose an approach to operationalizing core concepts needed to address interdependence in support of adaptive teamwork. We describe a formalized structure, joint activity graphs, built on interdependence design principles to capture the essence of joint activity. We describe the runtime requirements needed to dynamically exploit joint activity graphs and to support intelligent coordination during execution. We demonstrate the effectiveness of such a structure at supporting adaptability using the Capture-the-Flag domain with heterogeneous teams of unmanned aerial vehicles and unmanned ground systems. In this dynamic adversarial domain, we show how agents can make use of the information provided by joint activity graphs to generally and pragmatically react and adapt to perturbations in the joint activity, the environment, or the team and explicitly manage and exploit interdependence to produce effective teamwork. In doing so, we demonstrate how flexible and adaptive teamwork can be achieved through formally guided design that supports effective management of interdependence.

Intuiting machine failures

A key ingredient of effective collaborations is knowing the strengths and weaknesses of one’s collaborators. But what if one’s collaborator is a machine-classification system, of the sort increasingly appearing in semi-autonomous vehicles, radiologists’ offices, and other contexts in which visual processing is automated? Such systems have both remarkable strengths (including superhuman performance at certain classification tasks) and striking weaknesses (including susceptibility to bizarre misclassifications); can naive subjects intuit when such systems will succeed or fail? Here, five experiments (N=900) investigate whether humans can anticipate when machines will misclassify natural images. E1 showed subjects two natural images: one which reliably elicits misclassifications from multiple state-of-the-art Convolutional Neural Networks, and another image which reliably elicits correct classifications. We found that subjects could predict which image was misclassified on a majority of trials. E2 and E3 showed that subjects are sensitive to the nature of such misclassifications; subjects’ performance was better when they were told what the misclassification was (but not which image received it), and worse when the label shown was from another, randomly-chosen category. Crucially, in E4, we asked subjects to either (a) choose which image they thought was misclassified, or (b) choose the image that is the worst example of its category. While both instructions resulted in subjects choosing misclassified images above chance, subjects who were instructed to identify misclassifications performed better. In other words, humans appreciate the cues that mislead machines, beyond simply considering the prototypicality of an image. Lastly, E5 explored more naturalistic settings. Here, instead of an NAFC choice, subjects identified potential misclassifications from a stream of individual images; even in this setting, subjects were remarkably successful in anticipating machine errors. Thus, humans can anticipate when and how their machine collaborators succeed and fail, a skill that may be of great value to human-machine teams.

What factors may influence decision-making in the operation of Maritime autonomous surface ships? A systematic review

Maritime Autonomous Surface Ships (MASS) are being introduced with high levels of autonomy. This means that not only the role of the human operator is changing but also the way in which operations are performed. MASS are expected to be uncrewed platforms that are operated from a Remote Control Centre (RCC). As such, the concept of human-machine teams is important. The operation of MASS however shares some resemblance to other uncrewed platforms (e.g. Uncrewed Aerial Vehicles; UAVs) and it may be that lessons could be learned in relation to the development of MASS systems. A systematic literature review has been conducted, focusing on decision making, to generate insights into operating uncrewed vehicle platforms and design recommendations for MASS. Seven themes were revealed: decision support systems, trust, transparency, teams, task/role allocation, accountability and situation awareness. A Network Model was developed to show the interconnections between these themes which was then applied to a case study of an UAV accident. The purpose of doing this was to demonstrate the utility of the model to real-world scenarios and showed how each theme applied within the human-machine team.

Trust and communication in human-machine teaming

Intelligent highly-automated systems (HASs) are increasingly being created and deployed at scale with a broad range of purposes and operational environments. In uncertain or safety-critical environments, HASs are frequently designed to seamlessly co-operate with humans, thus, forming human-machine teams (HMTs) to achieve collective goals. Trust plays an important role in this dynamic: humans need to be able to develop an appropriate level of trust in their HAS teammate(s) to form an HMT capable of safely and effectively working towards goal completion. Using Autonomous Ground Vehicles (AGVs) as an example of an HAS used in dynamic social contexts, we explore interdependent teaming and communication between humans and AGVs in different contexts and examine the role of trust and communication in these teams. Drawing on lessons from the AGV example for the design of an HAS used for an HMT more broadly, we argue that trust is experienced and built differently in different contexts, necessitating context-specific approaches to designing for trust in such systems.