Human-centered AI Research Articles

Transitioning to Human Interaction with AI Systems: New Challenges and Opportunities for HCI Professionals to Enable Human-Centered AI

While AI has benefited humans, it may also harm humans if not appropriately developed. The priority of current HCI work should focus on transiting from conventional human interaction with non-AI computing systems to interaction with AI systems. We conducted a high-level literature review and a holistic analysis of current work in developing AI systems from an HCI perspective. Our review and analysis highlight the new changes introduced by AI technology and the new challenges that HCI professionals face when applying the human-centered AI (HCAI) approach in the development of AI systems. We also identified seven main issues in human interaction with AI systems, which HCI professionals did not encounter when developing non-AI computing systems. To further enable the implementation of the HCAI approach, we identified new HCI opportunities tied to specific HCAI-driven design goals to guide HCI professionals addressing these new issues. Finally, our assessment of current HCI methods shows the limitations of these methods in support of developing HCAI systems. We propose the alternative methods that can help overcome these limitations and effectively help HCI professionals apply the HCAI approach to the development of AI systems. We also offer strategic recommendation for HCI professionals to effectively influence the development of AI systems with the HCAI approach, eventually developing HCAI systems.

Keeping the organization in the loop: a socio-technical extension of human-centered artificial intelligence

The human-centered AI approach posits a future in which the work done by humans and machines will become ever more interactive and integrated. This article takes human-centered AI one step further. It argues that the integration of human and machine intelligence is achievable only if human organizations—not just individual human workers—are kept “in the loop.” We support this argument with evidence of two case studies in the area of predictive maintenance, by which we show how organizational practices are needed and shape the use of AI/ML. Specifically, organizational processes and outputs such as decision-making workflows, etc. directly influence how AI/ML affects the workplace, and they are crucial for answering our first and second research questions, which address the pre-conditions for keeping humans in the loop and for supporting continuous and reliable functioning of AI-based socio-technical processes. From the empirical cases, we extrapolate a concept of “keeping the organization in the loop” that integrates four different kinds of loops: AI use, AI customization, AI-supported original tasks, and taking contextual changes into account. The analysis culminates in a systematic framework of keeping the organization in the loop look based on interacting organizational practices.

Human-centered AI and robotics

Robotics has a special place in AI as robots are connected to the real world and robots increasingly appear in humans everyday environment, from home to industry. Apart from cases were robots are expected to completely replace them, humans will largely benefit from real interactions with such robots. This is not only true for complex interaction scenarios like robots serving as guides, companions or members in a team, but also for more predefined functions like autonomous transport of people or goods. More and more, robots need suitable interfaces to interact with humans in a way that humans feel comfortable and that takes into account the need for a certain transparency about actions taken. The paper describes the requirements and state-of-the-art for a human-centered robotics research and development, including verbal and non-verbal interaction, understanding and learning from each other, as well as ethical questions that have to be dealt with if robots will be included in our everyday environment, influencing human life and societies.

Fine-grained Human Analysis under Occlusions and Perspective Constraints in Multimedia Surveillance

Human detection in the wild is a research topic of paramount importance in computer vision, and it is the starting step for designing intelligent systems oriented to human interaction that work in complete autonomy. To achieve this goal, computer vision and machine learning should aim at superhuman capabilities. In this work, we address the problem of fine-grained human analysis under occlusions and perspective constraints. More specifically, we discuss some issues and some possible solutions to effectively detect people using pose estimation methods and to detect humans under occlusions both in the two-dimensional (2D) image plane and in the 3D space exploiting single monocular cameras. Dealing with occlusion can be done at the joint level or pixel level: We discuss two different solutions, the former based on a supervised neural network architecture for detecting occluded joints and the latter based on a semi-supervised specialized GAN that exploits both appearance and human shape attributes to determine the missing parts of the visible shape. To deal with perspective constraints, we further discuss a neural approach based on a double architecture that learns to create an optimal neural representation, which is useful to reconstruct the 3D position of human keypoints starting with simple RGB images. All these approaches have a critical point in common: the need for large annotated datasets. To have large, fair, consistent, transparent, and ethical datasets, we propose the adoption of synthetic datasets as, for example, JTA and MOTSynth. In this article, we discuss the pros and cons of using synthetic datasets while tackling several human-centered AI issues with respect to European GDPR rules for privacy. We further explore and discuss an application in the field of risk assessment by space occupancy estimation during the COVID-19 pandemic called Inter-Homines.

Documentation Matters: Human-Centered AI System to Assist Data Science Code Documentation in Computational Notebooks

Computational notebooks allow data scientists to express their ideas through a combination of code and documentation. However, data scientists often pay attention only to the code, and neglect creating or updating their documentation during quick iterations. Inspired by human documentation practices learned from 80 highly-voted Kaggle notebooks, we design and implement Themisto, an automated documentation generation system to explore how human-centered AI systems can support human data scientists in the machine learning code documentation scenario. Themisto facilitates the creation of documentation via three approaches: a deep-learning-based approach to generate documentation for source code, a query-based approach to retrieve online API documentation for source code, and a user prompt approach to nudge users to write documentation. We evaluated Themisto in a within-subjects experiment with 24 data science practitioners, and found that automated documentation generation techniques reduced the time for writing documentation, reminded participants to document code they would have ignored, and improved participants’ satisfaction with their computational notebook.

Reconsideration on human centered AI for autonomous vehicle

Reconsideration on human centered AI for autonomous vehicle

Epistemic Injustice as a Philosophical Conception for Considering Fairness and Diversity in Human-centered AI Principles

Epistemic Injustice as a Philosophical Conception for Considering Fairness and Diversity in Human-centered AI Principles

Interleaving Automatic Segmentation and Expert Opinion for Retinal Conditions.

Optical coherence tomography (OCT) has become the leading diagnostic tool in modern ophthalmology. We are interested here in developing a support tool for the segmentation of retina layers. The proposed method relies on graph theory and geodesic distance. As each retina layer is characterised by different features, the proposed method interleaves various gradients during detection, such as horizontal and vertical gradients or open-closed gradients. The method was tested on a dataset of 750 OCT B-Scan Spectralis provided by the Ophthalmology Department of the County Emergency Hospital Cluj-Napoca. The method has smaller signed error on layers B1, B7 and B8, with the highest value of 0.43 pixels. The average value of signed error on all layers is −1.99 ± 1.14 px. The average value for mean absolute error is 2.60 ± 0.95 px. Since the target is a support tool for the human agent, the ophthalmologist can intervene after each automatic step. Human intervention includes validation or fine tuning of the automatic segmentation. In line with design criteria advocated by explainable artificial intelligence (XAI) and human-centered AI, this approach gives more control and transparency as well as more of a global perspective on the segmentation process.

Creative AI and Musicking Robots.

This article discusses the creative and technical approaches in a performative robot project called “Embodied Musicking Robots” (2018–present). The core approach of this project is human-centered AI (HC-AI) which focuses on the design, development, and deployment of intelligent systems that cooperate with humans in real time in a “deep and meaningful way.” 1 This project applies this goal as a central philosophy from which the concepts of creative AI and experiential learning are developed. At the center of this discussion is the articulation of a shift in thinking of what constitutes creative AI and new HC-AI forms of computational learning from inside the flow of the shared experience between robots and humans. The central case study (EMRv1) investigates the technical solutions and artistic potential of AI-driven robots co-creating with an improvising human musician (the author) in real time. This project is ongoing, currently at v4, with limited conclusions; other than this, the approach can be felt to be cooperative but requires further investigation.

Towards Human-Centered AI: Psychological concepts as foundation for empirical XAI research

Abstract Human-Centered AI is a widely requested goal for AI applications. To reach this is explainable AI promises to help humans to understand the inner workings and decisions of AI systems. While different XAI techniques have been developed to shed light on AI systems, it is still unclear how end-users with no experience in machine learning perceive these. Psychological concepts like trust, mental models, and self-efficacy can serve as instruments to evaluate XAI approaches in empirical studies with end-users. First results in applications for education, healthcare, and industry suggest that one XAI does not fit all. Instead, the design of XAI has to consider user needs, personal background, and the specific task of the AI system.

Humans “in the Loop”?

Abstract More and more scholarly attention is being paid to the challenges of governing artificial intelligence and emergent technologies. Most of the focus remains on questions of how to preserve the human-centeredness of increasingly advancing machine-driven technologies. I problematize discourses of “human-centered AI” that prioritize human control over nonhuman intelligences as a solution for the challenges posed by emergent technologies like artificial intelligence. Posthumanism provides a compelling theoretical basis for this line of questioning and for reimagining alternative ethical constructs. I outline and consider three distinct scenarios in which (a) humans are at the center of command and control, (b) humans and nonhumans share control, (c) human oversight is completely removed. I suggest that more attention could be given to critical and speculative ways of reimagining the concepts of “human,” “nonhuman,” and human/nonhuman relations.

Humans “in the Loop”?

More and more scholarly attention is being paid to the challenges of governing artificial intelligence and emergent technologies. Most of the focus remains on questions of how to preserve the human-centeredness of increasingly advancing machine-driven technologies. I problematize discourses of “human-centered AI” that prioritize human control over nonhuman intelligences as a solution for the challenges posed by emergent technologies like artificial intelligence. Posthumanism provides a compelling theoretical basis for this line of questioning and for reimagining alternative ethical constructs. I outline and consider three distinct scenarios in which (a) humans are at the center of command and control, (b) humans and nonhumans share control, (c) human oversight is completely removed. I suggest that more attention could be given to critical and speculative ways of reimagining the concepts of “human,” “nonhuman,” and human/nonhuman relations.

Bridging the Gap Between Ethics and Practice

This article attempts to bridge the gap between widely discussed ethical principles of Human-centered AI (HCAI) and practical steps for effective governance. Since HCAI systems are developed and implemented in multiple organizational structures, I propose 15 recommendations at three levels of governance: team, organization, and industry. The recommendations are intended to increase the reliability, safety, and trustworthiness of HCAI systems: (1) reliable systems based on sound software engineering practices, (2) safety culture through business management strategies, and (3) trustworthy certification by independent oversight. Software engineering practices within teams include audit trails to enable analysis of failures, software engineering workflows, verification and validation testing, bias testing to enhance fairness, and explainable user interfaces. The safety culture within organizations comes from management strategies that include leadership commitment to safety, hiring and training oriented to safety, extensive reporting of failures and near misses, internal review boards for problems and future plans, and alignment with industry standard practices. The trustworthiness certification comes from industry-wide efforts that include government interventions and regulation, accounting firms conducting external audits, insurance companies compensating for failures, non-governmental and civil society organizations advancing design principles, and professional organizations and research institutes developing standards, policies, and novel ideas. The larger goal of effective governance is to limit the dangers and increase the benefits of HCAI to individuals, organizations, and society.

Human centered AI design for clinical monitoring and data management

Abstract Background In clinical settings, significant resources are spent on data collection and monitoring patients' health parameters to improve decision-making and provide better care. With increased digitization, the healthcare sector is shifting towards implementing digital technologies for data management and in administration. New technologies offer better treatment opportunities and streamline clinical workflow, but the complexity can cause ineffectiveness, frustration, and errors. To address this, we believe digital solutions alone are not sufficient. Therefore, we take a human-centred design approach for AI development, and apply systems engineering methods to identify system leverage points. We demonstrate how automation enables monitoring clinical parameters, using existing non-intrusive sensor technology, resulting in more resources toward patient care. Furthermore, we provide a framework on digitization of clinical data for integration with data management. Methods Activities of Daily Living (ADLs) are essential parameters, necessary for evaluating patients in mental health wards. Ideally logging the parameters should take place at hourly intervals; however, time constraints and lack of resources restrict the nursing staff to consolidating the overall impression during the day, relying on what they recall. Using design methods, sensors (e.g. infrared, proximity, pressure) are used to automate the acquisition of data for machine learning that correspond to the ADLs, considering privacy and other medical requirements. Results We present a concept of a room with sensors that can be deployed in clinical settings. Sensor data log ADLs, and provide machine learning data. A theoretical framework demonstrates how collected data can be used in electronic/medical health records. Conclusions Data acquisition of the ADLs with automation enable variable specificity and sensitivity on-demand. It further facilitates interoperability and provides data for machine learning. Key messages Our research demonstrates automated data acquisition techniques for clinical monitoring. Human centered AI design approach enables on-demand analysis of ADLs for mental health treatment.

Artificial Intelligence and Innovation Ethics

This symposium examines key questions posed by teaching ethics to artificial intelligence for business settings. A general question is how to balance the benefits and risks of AI, which is a significant concern with technological change. That concern is made more severe by the large-scale implications of AI on human life, including our understanding of what it is to be a human being and what entities can be properly treated as right holders. More specifically, several topics arise in the intersection of AI and Ethics that this panel will address. Fairness in the use of AI for business: When AI is used at a large scale for business, there is always a concern that it may also lead to drastic and large scale discrimination against some groups of people. For example, deep-learning systems may deny mortgage loans to members of certain groups when others with comparable financial resources receive loans, and this may occur even if none of the training data indicate group membership. It is thus crucial to design tools that can monitor the AI’s performance to continuously test for bias. A second crucial goal is to design methods to mitigate any such biases to the maximum extent possible. This research direction will involve both making fundamental contributions to AI and statistics in terms of developing these tools, and impactful use in business in many applications. A large ethics literature has carefully analyzed concepts of fairness, and this body of thought can be applied to AI. Many statistical measures of bias have been proposed, some of which are inconsistent with others. An ethical analysis can help evaluate whether measures have normative justification. Ethically grounded value alignment: Deep learning systems are frequently designed to reflect human values so as to avoid recommending decisions inconsistent with these values. Values are typically ascertained, however, in the much the same empirical way as facts and predictions - in this case, by analyzing large datasets that reflect human beliefs and preferences. Yet the AI community is coming to realize that a purely empirical approach can reflect biases and prejudices as well as acceptable moral values. There is no substitute for grounding value alignment in ethical principles that are independently derived, a manoeuvre that avoids the philosophically famous “naturalistic fallacy” of deriving ethical conclusions from purely factual premises. The deontological tradition in ethics provides the intellectual resources to develop rigorously defined and grounded principles that can be used to screen training sets or otherwise direct learning procedures. Human-Centered Explainable AI (XAI): Many industry experts have pointed out the critical need for human oriented explanation by AI systems. According to an IBM survey, about 60% of 5,000 executives were concerned “about being able to explain how AI is using data and making decisions.” However, the most successful algorithms in use today are not transparent. All of these models are fundamentally “black boxes” that include many layers of complex, typically nonlinear, transformations of inputs. It can be quite difficult for anyone to understand the algorithm’s output and/or why the model makes key predictions. Given these challenges, efforts to develop more interpretable, explainable, or intelligible algorithms comprise a key area of current research. The explainability of an algorithm plays a key role in detecting, enabling, and improving auditability, fairness, trust, and reliability. However, the definition of interpretability and desiderata of what makes a good explanation remain elusive and different researchers use different, often problem- or domain-specific, definitions. More alarmingly, this XAI research rarely involves systematic investigation of human responses with regard to a “What is a good explanation for machine learning output?” AI generates a variety of ethical questions at three interconnected levels. The first is the legal dimension: what laws should be enacted to govern AI? Should some particular aspect of AI be subject to legal regulation at all? Do we need to fashion specific legislation to address AI issues or rely on more general legal standards? The second is the social dimension, which raises questions about the social morality that should be cultivated concerning AI. What sort of culture will develop in response to AI? A third level is concerned with issues that arise for individuals and associations in their engagement with AI. That connects with corporations and associations, which still need to exercise their own moral judgment.

ARTIFICIAL INTELLIGENCE IN ENVIRONMENTAL MONITORING AND CONSERVATION

This paper presents and explores how to improve environmental sustainability with the help of artificial intelligence (AI). This study evaluates the potential environmental benefits of ai adoption, such as reductions in climate change, agriculture, ocean health, water resources, weather forecasting, and disaster resiliency. Upon conducting a thorough analysis of the available literature, it was found that there is a research deficiency in the utilization of artificial intelligence and decision support systems, as well as optimization models. This study employs qualitative analysis to explore sustainable uses of ai and the environmental impact of ai. It emphasizes. The importance of promoting environmental sustainability through ai and proposes “environmental sustainability. By ai" approach as a prerequisite for developing transparent, responsible, and human-centered ai systems. The study primarily focuses on identifying ways in which ai can be utilized for sustainable environmental practices, with a particular emphasis on the role of ai in promoting environmental sustainability. This study has some limitations, one of which is its limited scope. The paper does not provide an extensive analysis of global environmental policies, which could potentially identify areas for cooperation or common ground. Additionally, the study's focus on environmental sustainability means that it neglects the economic and social aspects of sustainability, which could be further explored. The study can aid stakeholders in comprehending global efforts to enhance environmental sustainability through the implementation of ai.

The Role of HCI in the Age of AI

ABSTRACTThis article examines some of the mystique surrounding AI, including the interrelated notions of explainability and complexity, and argues that these notions suggest that designing human-centered AI is difficult. It explains how, once these are put aside, an HCI perspective can help define interaction between AI and users that can enhance rather than substitute one important aspect of human life: creativity. Key to developing such creative interactions are abstractions and grammars of action and other notions; the article explores the history of these in HCI and how they are to be used in the contemporary interaction and design space, in relation to AI. The article is programmatic rather than empirical though its argument uses real-world examples.

Toward human-centered AI

column Share on Toward human-centered AI: a perspective from human-computer interaction Author: Wei Xu Intel Corporation Intel CorporationView Profile Authors Info & Claims InteractionsVolume 26Issue 4July-August 2019 pp 42–46https://doi.org/10.1145/3328485Published:26 June 2019Publication History 100citation13,831DownloadsMetricsTotal Citations100Total Downloads13,831Last 12 Months3,963Last 6 weeks402 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access

Guiding computers, robots to see and think

Fei-Fei Li, co-director of Stanford University's Human-Centered AI Institute, wants to create algorithms that can learn the way human babies do.