Transparency is a critical component when building artificial intelligence (AI) decision-support tools, especially for contexts in which AI outputs impact people or policy. Effectively identifying and addressing user transparency needs in practice remains a challenge. While a number of guidelines and processes for identifying transparency needs have emerged, existing methods tend to approach need-finding with a limited focus that centers around a narrow set of stakeholders and transparency techniques. To broaden this perspective, we employ numerous need-finding methods to investigate transparency mechanisms in a widely deployed AI-decision support tool developed by a wildlife conservation non-profit. Throughout our 5-month case study, we conducted need-finding through semi-structured interviews with end-users, analysis of the tool’s community forum, experiments with their ML model, and analysis of training documents created by end-users. We also held regular meetings with the tool’s product and machine learning teams. By approaching transparency need-finding from a broad lens, we uncover insights into end-users’ transparency needs as well as unexpected uses and challenges with current transparency mechanisms. Our study is one of the first to incorporate such diverse perspectives to reveal an unbiased and rich view of transparency needs. Lastly, we offer the FAccT community recommendations on broadening transparency need-finding approaches, contributing to the evolving field of transparency research.

To enable the development and use of safe and equitable artificial intelligence (AI) systems, AI engineers must monitor deployed AI systems and learn from past AI incidents where failures have occurred. Around the world, public databases for cataloging AI systems and resulting harms are instrumental in promoting awareness of potential AI harms among policymakers, researchers, and the public. However, despite growing recognition of the potential of AI systems to produce harms, causes of AI systems failure remain elusive and AI incidents continue to occur. For example, incidents of AI bias are frequently reported and discussed, yet biased systems continue to be developed and deployed. This raises the question – how are we learning from documented incidents? What information do we need to analyze AI incidents and develop new AI engineering best practices? This paper examines reporting techniques from a variety of AI stakeholders and across different industries, identifies requirements towards the design of effective AI incident documentation, and proposes policy recommendations for augmenting current practice.

AbstractDesign education increasingly blends technology learning with sociotechnical challenges, but little is understood about how students simultaneously engage with both of these elements. In this preliminary study, we describe the results of two offerings of a design course focusing on disaster response at a major public research institution. We present a preliminary analysis of 52 students’ course reflections suggesting that sociotechnical challenges uniquely contextualize technology during project-based learning, presenting promising opportunities for future design education and research study.

Current test and evaluation (T&E) methods for assessing machine learning (ML) system performance often rely on incomplete metrics. Testing is additionally often siloed from the other phases of the ML system lifecycle. Research investigating cross-domain approaches to ML T&E is needed to drive the state of the art forward and to build an Artificial Intelligence (AI) engineering discipline. This paper advocates for a robust, integrated approach to testing by outlining six key questions for guiding a holistic T&E strategy.

We identify three specific areas of focus to advance Robust and Secure AI for defense and national security:• Improving the robustness of AI components and systems and the need to go beyond accuracy measurements to capture achievement of mission outcomes• Designing for security challenges in modern AI systems including novel attack surfaces and patterns, as well as strategies for risk mitigation• Developing processes and tools for testing, evaluating, and analyzing AI systems and adoption of comprehensive test and evaluation approachesFor each area, we identify ongoing work as well as challenges and opportunities in developing and deploying AI systems with confidence.

Engineering graduates must know how to frame and solve non-routine problems. While design classes explicitly teach problem framing and solving, it is lacking throughout much of the rest of the engineering curriculum and is often relegated to capstone classes at the end of the students' educational experience. This paper explores problem framing and solving through the lens of experiential learning theory. It captures core problem framing and solving approaches from critical, design and systems thinking and concludes with a table of learning outcomes that might be drawn upon in designing an engineering curriculum that more fully develops the problem framing and solving capabilities of its students.

Business schools around the world are increasingly integrating design thinking into their curricula, which invites definition of a new professional identity for individuals effectively working at the intersection of the two fields. Research is needed, however, to understand what such a profession might entail and in what ways students might best be prepared to enter it. Defining the capabilities that a design thinking business professional will embody requires an understanding of the ways design thinking complements or contradicts more traditional critical thinking approaches applied to business decisions. This chapter adopts a framework grounded in learning theory to describe the four basic capabilities needed to frame and solve problems in practice: observing and noticing; framing and reframing; imagining and designing; and making and experimenting. Using that framework, it unpacks the contributions that design and critical thinking make to the development of each of the capabilities in a business setting and highlights challenges associated with helping business students learn design thinking and how they might use it to complement other approaches to framing and solving business problems.

Design education continues to interest students across universities today. Students choose among myriad of programs that provide pathways to becoming a professional designer, but do so through varied courses of study that depend upon factors such as the school, the program's home discipline, the student population, and the intended professional design field. When creating a multidisciplinary design certificate at the University of California, Berkeley that would allow students to take design courses from multiple fields including business, engineering, humanities and environmental design, faculty from the disciplines engaged in a series of discussions about programmatic learning outcomes. As a result, a question arose about the extent to which disciplinary perspective influences design pedagogy and the resulting training provided to the student. This paper examines learning outcomes associated with design programs anchored in four disciplines and assesses similarities and differences amongst them. In doing so, it aims to uncover ways in which disciplinary focus influences programmatic goals and to identify implications of differences among them for the construction of multidisciplinary design programs.

A framework for quantifying environmental impact in future energy codes is explored based on exergy analysis. An example exergy analysis is performed that quantifies the environmental impact of electric resistance heating with exergetic efficiency of only 13% compared to 30% and 40% for air source and ground source heat pumps. This provides motivation for removing electric resistance as a space heating option in future energy codes in the Northwest.

The market for solid-state lighting (SSL) systems has expanded 40-fold in installed lamps since 2001. At the same time, systems which preserve materials over time and promote material reuse are getting increasing attention in light of calls for reducing consumption of natural resources. As new lighting technology products are designed and brought to market, consideration must be given to how products will be managed throughout the life-cycle as well as their end-of-life (EOL) fate. Lighting-as-a-service (LaaS) business models have emerged as a potential strategy for preserving the materials embedded in lighting products. In this paper, we examine the cost and environmental implications of technology management decisions in the context of the street lighting industry, employing life-cycle assessment and a Markov Decision Process model. The goal of the research is to determine a policy that minimizes expected costs and emissions for the system over a fixed time horizon thus reducing uncertainty for managers. The model used in the paper evaluates the optimal replacement strategies for street lighting products and additionally connects the result to the optimal EOL product trajectory, taking both costs and carbon emissions into account. In doing so, we are able to more deeply understand the role that LaaS business models might play in enabling closed-loop systems within the street lighting industry.

Technological innovation has the promise to catalyze reduced environmental impacts through higher energy efficiency, product design that promotes reuse, and use of benign materials. However, as new technologies emerge, it is easy to lose sight of the environmental burden associated with technology transition and premature replacement of technology. Though product turnover can lead to gains in energy efficiency, it also can contribute to the creation of expanded and diversified waste streams. The market for residential lighting products continues to rapidly evolve, with new LED products emerging each year. Technology change in the lighting industry has unintended consequences, including new waste streams associated with the obsolete products. This paper examines the waste burden that occurs as a consequence of technology change when lighting products are replaced or upgraded. To understand the impact of residential lighting replacement, the authors used life cycle assessment to evaluate LED products. We defined three personas representing various typical technology adopters within the product adoption curve to quantify the volume and change rate of lighting products. The personas have also been used to estimate the waste impact for larger populations, and the results show that while an innovator's waste burden is lowest amongst the three personas, a non-trivial waste burden accompanies the transition to LED lighting products for all three personas. The results quantify the waste burden of high performance lighting and further motivate the development and implementation of recycling programs and policies to prevent waste diverted to landfills by consumers. The results also confirm that attention must be paid to how to reduce the waste burden of LED lighting products through improved design and lighting as a service models.

Recent technological advances have made light emitting diode (LED) lamps available to the domestic replacement lightbulb market at much lower prices than ever before. Longer life and higher efficiency have occurred as the products matured. This research studied how LED lamps have changed in the last 5 years and how changes to the lamps impact the environment. A life cycle assessment has been conducted using data from the Ecoinvent database to determine how three recently manufactured lamps compare to the 2011 counterpart in terms of environmental impacts. We found that the newer lamps all perform better than the 2011 A19 LED on environmental impacts and all are now significantly better than traditional lighting technologies. All three lamps have improved in key environmental impact areas, specifically hazardous waste to landfill due to reduced use of metals. The LED lamps studied are outperforming older LED lamps in all environmental impact categories.

At the University of California at Berkeley, the vision of a 2020 Engineer with the skills to solve the global challenges of today is being realized through an emergent program focused on design for social impact. A minor for Ph.D. students, Development Engineering (DevEng) is a degree program and accompanying ecosystem that aims ". . . to create technology interventions in accordance with the needs and wants of individuals living within complex, low-resource settings.'' This paper explores how the competencies conveyed through the DevEng program overlap with and go beyond the criteria laid out for the Engineer of 2020. Today's engineering student engages with a wide variety of global problems, only some of which are addressed by solely technological solutions. Academic programs across the country, including DevEng, have recognized that a unique skillset is needed to address these complex challenges. As a result, there is a growing collection of academic programs (e.g., DevEng, Design for Social Impact, Humanitarian Engineering) that train technologically adept engineers to work in multifunctional teams while attempting to have a positive social impact on the world. In this paper, the authors examine the skill development activities that students in the DevEng Program are exposed to and whether they meet or exceed the goals and principles of the 2020 Engineer. The paper further describes the DevEng learning objectives and accompanying ecosystem opportunities to determine whether the program is on track in providing students with a multidisciplinary set of skills and diverse experiences that effectively train the Engineer of 2020.

Consumers in the US market and across the globe are beginning to widely adopt light emitting diode (LED) lighting products while the technology continues to undergo significant changes. While LED products are evolving to consume less energy, they are also more complex than traditional lighting products with a higher number of parts and a larger number of electronic components. Enthusiasm around the efficiency and long expected life span of LED lighting products is valid, but research to optimize product characteristics and design is needed. This study seeks to address that gap by characterizing LED lighting products' suitability for end of life (EOL) recycling and disposal. The authors disassembled and assessed 17 different lighting products to understand how designs differ between brands and manufacture year. Products were evaluated based on six parameters to quantify the design. The analysis indicates that while the efficiency of LED products has improved dramatically in the recent past, product designers and manufacturers could incorporate design strategies to improve environmental performance of lighting products at end-of-life.

Workplace and classroom culture, persistent stereotypes, and gender-specific pressures have all been suggested as reasons for the gender discrepancy in engineering. This article shares some of the behind-the-scenes observations and strategies that have proven successful to recruit, engage, retain, and propel women on their engineering career trajectories. Though the HESE program is small when comp...

Academic programs focused on engineering entrepreneurship are growing in number and popularity at American universities. However, the fields of engineering, entrepreneurship and technology-based entrepreneurship struggle to recruit and retain female students: a historic and endemic failure at obtaining gender-balanced participation. Understanding the motivations of women engaged in such academic programs can influence recruitment and retention strategies and lead to the creation of transformative learning environments and entrepreneurial ecosystems. This study investigates the motivations that women have for participating in the Humanitarian Engineering and Social Entrepreneurship (HESE) program at Penn State. In this five-year study, females from diverse academic majors cited the following to be their top motivations for participating: a) becoming a global professional; b) making a difference; and c) applying theory learned in school to address a problem. In contrast, males cited 1) making a difference 2) participating in an exciting real-world project and 3) applying theory to address a problem, as their top motivations. Studying the motivations and perceptions of enterprising women and men provides compelling insights to help make academic engineering entrepreneurship programs more appealing to both female and male students.

Manufacturing affects all three dimensions of sustainability: economy, environment, and society. This paper addresses the last of these dimensions. It explores social impacts identified by national level social indicators, frameworks, and principles. The effects of manufacturing on social performance are framed for different stakeholder groups with associated social needs. Methodology development as well as various challenges for social life cycle assessment (S-LCA) are further examined. Efforts to integrate social and another dimension of sustainability are considered, with attention to globalization challenges, including offshoring and reshoring. The paper concludes with a summary of key takeaways and promising directions for future work.

Engineering is one of the most global professions, with design teams developing technologies for an increasingly interconnected and borderless world. In order for engineering students to be proficient in creating viable solutions to the challenges faced by diverse populations, they must receive an experiential education in rigorous engineering design processes as well as identify the needs of customers living in communities radically different from their own. Acquainting students with the unique context and constraints of developing countries is difficult because of the breadth of pertinent considerations and the time constraints of academic semesters. This article describes a tool called Global Biomedical Device Design, or GloBDD, that facilitates simultaneous instruction in design methodology and global context considerations. GloBDD espouses an example-centric approach to educate students in the user-centered and context-driven design of biomedical devices. The tool employs real-world case studies to help students understand the importance of identifying external considerations early in the design process: issues like anthropometric, contextual, social, economic, and manufacturing considerations amongst many others. This article presents the rationale for the tool, its content and organization, and evaluation results from integration into a junior-level biomedical device design class. Results indicate that the tool engages students in design space exploration, leads them to making sound design decisions, and teaches them how to defend these decisions with a well-informed rationale.

Developing countries like Kenya suffer from a severe lack of healthcare professionals.Access to primary care is further limited when large distances must be travelled for basic medical care,and often times the decision to seek professional help is made too late. Preventive health services cansave lives by diagnosing serious medical conditions early and subsequently reducing the likelihood oflife-threatening diseases. The rapid increase in the global incidence of chronic diseases like diabetes,hypertension, and heart disease makes preventive health services particularly important. Trainedcommunity health workers (CHWs) operated the Mashavu telemedicine system in Nyeri Countyof Kenya over the course of fi ve months. The Mashavu system comprised of a netbook computer, aweb camera, and other basic supplies. CHWs provided patients with two alternative services: (1)body measurements including blood pressure, weight, and body mass index (BMI) or (2) a fullconsultation that supplemented the measurements with health indicators and a brief survey. Alightweight website was developed to allow CHWs to send this information to a nearby nurse. Thiswebsite was accessed by leveraging the 3G cell networks that are ubiquitous in Kenya. During thefi ve-month pilot, 777 patients paid for these services. Nearly one-third of patients used the serviceas an opportunity for a routine health check. The median age among patients was 40 years and 35 %of all patients were between the ages of 21-35. The pilot described in this paper demonstrated toCHWs that community members value health information and are willing to pay a small fee for it.This pilot suggests that as developing countries struggle to address the double burden of chronicand infectious disease, fee-for-service preventive health services can serve as an innovative meansto achieve economic development and improved community health.