Skip to main content

Currently Skimming:


Pages 29-71

The Chapter Skim interface presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter.
Select key terms on the right to highlight them within pages of the chapter.


From page 29...
... The instrumental values listed below are ones frequently raised in conjunction with computing research. Some expectations for responsible computing research are themselves an amalgamation of instrumental values.
From page 31...
... Computing research on system assurance similarly focuses on the value of providing reasons to expect that the system will behave appropriately, particularly in complex environments. • Explainability, interpretability, and intelligibility -- These concepts are grouped together as they have all been proposed as ways to promote understanding of increasingly complex computational systems, and thus to support meaning ful deliberation, oversight, or use of these systems.
From page 33...
... It is computing's participation in this social ecosystem and, consequently, its far-reaching societal impact that give rise to the ethical challenges and questions of responsibilities increasingly posed to computing research. The sociotechnical approach10 explained in this section provides an important framework for the computing research community in its pursuit of understanding ways to identify and address these challenges and calls for greater accountability.
From page 34...
... In particular, this approach draws social considerations into focus for even the most seemingly purely technical systems, and thereby shows the advantage of fram ing the challenges of responsible computing research as sociotechnical problems. More socially accountable technologies that support people's manifold values require the expertise of both social scientists and computing researchers.13 2.2.1 Sociotechnical Systems Briefly Explained The social contexts that participate in this feedback loop include the many interpersonal, linguistic, cultural, professional, institutional, and historical experiences that shape individuals as well as their personal experiences.
From page 35...
... As a result, seemingly purely technical systems are not just technical, and approaches to responsible computing must grapple with the myriad ways computing research interacts with people and the social contexts they inhabit. The facial recognition example presented below illuminates and explores the sociotechnical nature of computing technologies.
From page 36...
... Hundreds of computer researchers, using these collections, advanced the ability of software to identify any single face with a computational projection of the mathematical likelihood that an image taken in real time matched the face in front of it. Second, combining box 36 FOSTERING RESPONSIBLE COMPUTING RESEARCH
From page 37...
... They share group identities as well as linguistic and cultural traditions, and they navi gate a world of laws and economic conditions. Models of individual human cognition and behavior on which computing research often relies tend to flatten or disregard such social phenomena.
From page 39...
... Technology design can not fully be decoupled from the phenomena of social contexts. For example, the social phenomena of gender identity -- people's everyday experi ence of gender -- is constantly evolving with the macro level of legal requirements, the meso level of cultural norms, and the micro level of self-esteem.
From page 41...
... This loop creates a challenge for responsible computing research: social phenomena and their relationships with tech nologies can seem stochastic and hard to interpret. From the viewpoint of any individual person and any specific technical system, it can seem impossible to predict, let alone pre vent, what happens with technologies as they unfold.
From page 42...
... The sociotech nical perspective described in Section 2.2 along with ethical analyses of values and trade offs as described in Section 2.1, combined with methods of ethical reasoning and such social science methods as ethnographic observation, in-depth interviews, survey studies, and historical analysis, can support computing researchers in identifying and resolving the ethical and societal impact challenges that arise from introducing novel technologies into social worlds. Chapter 3 illustrates their use in identifying underlying roots of such challenges.
From page 43...
... Rather, responsible computing research requires that they collaborate with experts in other disciplines who can bring these important instruments to bear as computing research is designed and carried out. Theoretical Foundations from Ethical and Social Science Frameworks 43
From page 44...
... The chapter's discussion also makes evident the importance for computing research of incorporating into it consideration of ethical val ues and trade-offs, of methods from the social sciences, as described in Chapter 2, and of the multidisciplinary collaborations important to realizing ethically and societally respon sible technologies and avoiding potential negative consequences of novel technologies. The chapter thus lays a basis for appreciating the report's recommendations.
From page 45...
... . Some might object that computing researchers have no roles or responsibilities when it comes to downstream product design, implementation, or deployment.
From page 47...
... Similarly, one cannot conduct value-free or "valueuniform" computing research, and so cannot avoid deciding which values and interests will, for the purposes of this research, be prioritized. 3.1.2 Preserving Human Dignity The concept of human dignity is rooted in human intuitions and deeply held values as well as domestic and international law.
From page 48...
... A range of examples help illustrate connections between computing research and potential impacts -- both positive and negative -- on human dignity. On the positive side, adjustments to workplace scheduling software to give more control to workers can significantly improve workers' quality of life and sense of dignity.3 On the negative side, any strictly rule-based system in the administration of criminal justice will likely be both over- and under-inclusive, and so there need to be humans in the loop to ensure that all parties retain their human dignity.4 The double-edged potential of computing research for human dignity is readily seen in possible uses of automation to help handle the crush of civil litigation.5 On the one hand, many people currently lack realistic access to human expertise in situations where civil litigation might be appropriate for them, and so computing technology could help restore their autonomy and dignity.
From page 49...
... The absence of human discriminatory animus explicitly in these systems, however, in no way ensures that those systems will not reproduce, in digital form, structurally biased social phenomena, as critiques of such technologies as facial recognition and predictive policing show.8 Computing research avenues for addressing this challenge include more intentional participation of diverse users in training systems and development of objectives for model training that counter such biases. Predictions of large-scale social phenomena are vital contributors to public policy and are central to social scientific research.
From page 50...
... This fact does not imply that such models should not be developed, but that when they are developed or deployed, if at all, computing researchers, developers, and deployers should do so with full knowledge of the significant moral risks associated with predicting individual behavior in uses that make high stakes decisions about people's lives; researchers should therefore clearly delineate the limitations of the capabilities and the intended scope of applicability of their methods and systems. Second, computing systems are also being used to predict human behavior in order to calibrate interventions that are then used to shape that behavior.
From page 51...
... Computing may be able to help address these social problems, but to do so it needs to be designed to support people in roles that address those problems. Even so, there are deeper underlying problems beyond the sphere of computing research -- the litigiousness of a society, the large volume of small-claims disputes, and difficulties interpreting legal documents -- that need to be addressed.
From page 53...
... Lastly, Levy discussed how computing technology has recently been deployed to monitor the tone of call center workers and how such monitoring could be quite overbearing and create an unproductive and inhuman psychological burden on those workers. Computing technologies shape social relationships between manager and employee in these examples.17 One area where computing technologies have not easily integrated with existing institutional structures and norms is health care.
From page 54...
... To address environmental externalities, computing researchers will need to broaden how they view the impacts of computing technologies and consider the wide range of ethical values and trade-offs raised by these externalities. Significant strides have been made to enhance the energy efficiency of data centers and other infrastructure, drawing on computing research on more energy efficient components, architecture, and power and cooling designs as well as using renewable energy sources.
From page 55...
... By educating government policymakers, and monitoring uses of computing technologies, computing researchers can help society better manage these risks. Although current computing researchers are not responsible for continued use of obsolete computing technology (e.g., in air traffic control or nuclear weapons command and control)
From page 56...
... Insufficient awareness of and lack of attention to systemic racism and sexism -- which includes the implicit perpetuating of biases that were once explicit and blatant -- have raised and continue to raise ethical and societal impact concerns.22 Computing researchers and the computing research community at large cannot by themselves change these structures. Redress requires action at every level from society and government to individual researchers, and much collaborative work.
From page 57...
... Differences in race and gender are not simple differences, but structurally hierarchical ones that almost always yield power and importance differences. A lack of attention to structural racism and sexism have led to many current problems of computing technology, problems which often have origins in a similar lack of attention in the computing research that underlies those systems.
From page 58...
... Integrating Ethical and Societal Issues into Training Computing researchers are commonly trained in computer science, computer engineer ing, or a related discipline. Typical curricula and research training in computer science and engineering still involve little or no exposure to the social and behavioral sciences.
From page 59...
... Al 28 though computing ethics is an independent area of research in its own right,29,30 there is an ongoing need for more practically engaged and technically informed philosophical scholarship to help computing researchers better understand the ethical implications of their work. For responsible computing research, there is an urgent need for computing researchers to acknowledge that considering the societal and ethical implications of their work is an essential component of that work.
From page 61...
... The discussion and examples presented in the subsec tions show that responsible computing research requires computing researchers to be transparent about the intended use situations for the computing methods and artifacts their research produces, limitations in their power and applicability, the assumptions about people's capabilities their performance rests on, and the range of situations in which they have been tested. These responsibilities are reflected in Recommendation 6 (especially 6.5)
From page 63...
... As Kevin Fu observed in his remarks to the committee, the Bluetooth protocols in question were used in a wide array of medical devices.41 A large number of hospitals and medical offices had to be notified about the need to apply patches to protect patient safety and effectiveness. Such "open worlds problems" are of relevance to computing researchers whether they are developing systems for real-world deployment or developing methods that others may use for such purposes.
From page 65...
... Sources of Ethical Challenges and Societal Concerns 65
From page 66...
... For instance, one approach taken toward developing such systems is to consider the degree to which a person can consistently predict the outcomes of such a computing system; the higher the interpretability, the easier it is for a user to comprehend why certain decisions or predictions have been made.56 Computing research has only begun to address the need for transparency of these systems. 3.3 SOCIETAL CONTEXTS AND DESIGN AND DEPLOYMENT CHOICES Many of the adverse ethical and societal impacts of computing technology described in the preceding sections result from choices made during design or deployment of a 55 B
From page 67...
... 3.3.1 Ideation and Design Stage Failure to consider a full range of consequences early in the process of developing computing research increases the risk of adverse ethical or societal impacts because researchers have less time to find them and processes long under way (with a variety of investments in the research protocol already undertaken) are more difficult to reform or stop.
From page 69...
... Biased outcomes often result from sampling inadequacies, in particular when predictive functions are developed using data that happens to be available -- data gathered from what is called "convenience sampling" -- rather than data from carefully designed empirical work or curated with attention to the categories and types of data absent from the data set. Such bias and coverage inadequacies in training and benchmark data typically result from using data collected from online sources or preexisting collections of research study data from conveniently available study participants.
From page 71...
... 715-726 in Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society, Association for Computing Machinery, New York, https://doi.org/10.1145/3461702.3462628. Sources of Ethical Challenges and Societal Concerns 71


This material may be derived from roughly machine-read images, and so is provided only to facilitate research.
More information on Chapter Skim is available.