Education
and Information
1 101-122 (1996)
Technologies
A framework for implementing and teaching the social and ethical impact of computing C. DIANNE
MARTIN
Electrical Engineering and Computer Washington, DC, USA. E-mail:
[email protected] CHUCK
Science, George Washington
HUFF
Psychology Department, MN, USA. E-mail:
[email protected] DONALD
St Olaf College, 1520 St Olaf Avenue, Northfield,
GO’ITERBARN
Computer Science, East Tennessee State University, E-mail:
[email protected] KEITH
University,
Johnson
City, TN, USA.
MILLER
Computer Science, University E-mail:
[email protected]
of Illinois at Springfield,
Springfield,
IL, USA.
This article describes the work of Project ImpactCS, which was initiated to define the core content and pedagogical objectives for integrating social impact and ethics into the computer science (CS) curriculum. Over a three year period the project will address three major problems that hamper the implementation of across-the-board curriculum change: the lack of a well-specified definition of what the core content and learning objectives should include, the lack of a strategy for adapting and adopting existing materials that address the core topics into the existing CS curriculum, and the lack of awareness and expertise on the part of most CS faculties regarding the need and methodology for presenting such material in their courses. This report provides the conceptual framework and describes the learning objectives, called knowledge units, for defining a new content area in the computer science curriculum. It also discusses strategies and innovative pedagogical techniques for implementing the knowledge units into the curriculum. KEYWORDS: higher education; informatics; curriculum development; implications; social issues. 1360-2357
0 1996 Chapman
& Hall
102
Martin et al.
INTRODUCTION
Computer science has advanced rapidly in the last several decades, and this advance has necessitated the continual revision of the curriculum in an evolving discipline. One of the fundamental changes in computer science in the last decade is the realization that the context in which technology is used must be taken into account in its design, partly because of the ethical implications of its use, and partly because understanding the consequences of use helps to inform and improve the design (Shneiderman, 1990). This recognition was included as one of the foundational principles in Computing Curricula 1991 (ACM/IEEE, 1991), and has been a part of Computer Science Accreditation Board curriculum standards for over a decade (ACM Curriculum Committee on Computer Science, 1979; Computer Sciences Accreditation Board, 1987). Thus, as a part of the natural evolution of a maturing discipline, the charge to include the social, ethical and professional context of the discipline has been added in the core undergraduate curriculum.
Project ImpactCS The importance of grappling with these questions was underscored by Computing Curricula 1991 (ACM/IEEE, 1991) which calls the social, ethical and professional issues one of the subject areas in its common requirements for the undergraduate computer science curriculum. However, the report provides high-level guidelines, rather than specific recommendations for those who would teach this subject area. To this end Project ImpactCS was initiated with the goal to define the core content and pedagogical objectives for integrating social impact and ethics into the computer science curriculum. Over the course of three years the project will address three major problems that hamper the implementation of across-the-board curriculum change: the lack of a well specified definition of what the core content and learning objectives should include, the lack of a strategy for adapting and adopting existing materials that address the core topics into the existing CS curriculum, and the lack of awareness and expertise on the part of most CS faculties regarding the need and methodology for presenting such material in their courses. Using Computing Curricula 1991 report (ACM/IEEE, 1991) as a model, this report provides the conceptual framework and rationale for defining a new content area in the computer science curriculum with the same intellectual rigour that the other content areas are defined. This report represents the culmination of effort over the past two years to address the first problem, the development of a well-specified core of content and pedagogical objectives for integrating social impact and ethics into the required computer science curriculum. The work began at a two day meeting held in August 1994, with the convening of a Steering Committee comprised of 25 experts all of whom are well known in computer ethics and social impact from a number of different perspectives: philosophers, social scientists, ethicists and computer science professors with expertise in CS curriculum accreditation issues. From that meeting emerged the conceptual framework, the pedagogical objectives and a database of topics and teaching methodologies (lmpactCS Steering Committee, 1995).
Social and ethical impact of computing
103
During the second stage of the project, specific knowledge units and learning objectives were developed to enable the new subject area to be implemented into the teaching curriculum (ImpactCS Steering Committee, 1996). The final task of the project will be to gather materials and resources to be presented at a pilot faculty enhancement seminar to present the basic principles and skills and practical pedagogical techniques to computer science faculty members. DISCUSSION
Conceptual framework It is clear that the study of ethical and social issues in computing is interdisciplinary in nature. Ethicists from both philosophy and theology, historians, social analysts, sociologists, anthropologists and psychologists have all contributed heavily to the research in this area. Instead of suggesting the student learn each discipline separately, one suggests that from the perspective of computer science, every ethical concern is located at a particular level of social analysis. Only an analysis that takes account of at least three dimensions: the technical, the social and the ethical, can adequately represent the issues as they concern computer science in practice. Considering each dimension separately provides some insight, but it is only in their interaction that one can begin to grasp the complexity of the issues. A careful analysis of any ethical issue will need to specify and examine the ethical issues, the level of social analysis and the technical aspects of the issue at hand. The knowledge space defined by these dimensions is summarized in Fig. 1. The two dimensions shown in detail are the level of social analysis and the particular ethical issues that arise in technology. A third dimension is indicated, but not specified in detail in the table - the different technologies that require analysis from some portion of the two-dimensional space. Each of the ethical issues represented by a column in the table have been dealt with at great length in both primary and secondary scholarly literature, and in popular and academic venues. Philosophical work done on the concept of property could alone fill several bookshelves. Each of the levels of social analysis represented by the rows of the table also have a literature associated with them that includes thousands of references. Thus, the combination of these two dimensions results in such an overwhelming wealth of research and analysis that it could be difficult to determine where to start. Fortunately, there is a clear rule to help one determine the starting point. What topics, principles and skills from this array will be relevant to computer science students at the undergraduate level? Thus, a consideration of the issues that arise for computer professionals, and which are often dealt with in codes of ethics, is both a fundamental and an integral part of any topic covered in the table. Fundamental, in that the approach assumed when addressing these issues should be that of the computing professional and not that of either the philosopher or
Figure 1. The threedimensional
Global
Nations
Institutional Sectors
Cultures
Organizations
Communities & Groups
knowledge space
Quality of Life
Ethical Issues
pits of Ethical Analysis
Social and ethical impact of computing
105
the social scientist. Integral, because the issues of individual and professional responsibility should be explicitly addressed when dealing with every topic in the space (though for pedagogical purposes, a column with that title has been included).
The interaction of the dimensions As stressed earlier, each of the three analytical approaches represented by the three dimensions in Fig. 1 (ethics, social science, technology) is important in its own right. Each has its own practitioners, literature and approach. It is important to note that any ethical or social issue in computing will always be multidimensional and thus requires careful presentation. Some knowledge of these dimensions on the part of the instructor would be essential to the success of a course on social and ethical issues in computing. However, the important thing to note is that all three approaches are essential to understanding any particular issue. General principles about social analysis may be useful, but need to be connected with their implications for ethical practice in computer science. Ethical principles of argument need to be anchored in the uses and effects of particular technologies at particular levels of social analysis. Reviews of the various technologies available need to be conducted in light of the social situations and ethical concerns they raise. This multidimensionality is represented in Fig. 2. The shaded areas in Fig. 2 present a picture of hypothetical areas of social and ethical concern for four particular technologies. Note that each technology will have differing areas of concern, and that each area is described by both a level of social analysis and an ethical issue. Any particular issue, such as privacy in corporate records or risks in medical technology, will cover many levels of social analysis, several different ethical issues, and will be spread across differing implementations of the technology. A careful analysis of any issue must address all these dimensions. Even a short analysis like one found in a module in a course needs at least to make students aware of the complexity of what may seem to be a simple problem. Importance
of dimensions
In this section is presented a short explanation of the rows and columns of the charts used in Figs 1 and 2. The purpose here is simply to indicate why each of the levels of analysis or ethical issues identified in the chart is important to an analysis of computer technology from the viewpoint of the computer professional. Clearly, one cannot claim that the set of ethical issues and social analyses provided here are a comprehensive taxonomy of the field. That is too much to expect from a new approach, and only extensive use will determine if the categories or conceptual scheme need to be revised. One can, however, make the claim that this is a reasonably comprehensive and quite useful categorization of the issues in their ethical, social, and technological aspects, and that it provides an organizing framework that has been rare in most approaches to the ethical and social issues in computing.
106
Martin et al.
Electronic Communication Topics of Ethical Analysis Responsibiliiy
1
Ethical Issues
Computer Aided Manufacturing
Communities &
1 31 Organizations
IFI
Institutional
I
Medical Technolgy
Social and ethical impact of computing
107
Artificial Intelligence
Figure 2.
Social and ethical areas of concern for different technologies
Levels of social analysis A common approach in social analysis is first to determine the level of analysis required for a particular issue and then to apply the tools, literature and methods appropriate to that level. This is not to suggest that there is a ‘privileged’ level, since many issues must be analysed at several levels in order to understand them. But it is to suggest that determining the appropriate levels of analysis will help in teaching social and ethical issues in computing. In this section, one provides a description of each level of analysis, some examples of a few issues that can be profitably analysed at each level, and some suggestions about how the particular level interacts with the ethical issues represented by the columns in Fig. 1. Individuals. Even in relatively homogeneous groups or organizations, there will be individual differences in approach to technology and individual interests in the use of that technology. Some of these differences are based on physical and others on psychological differences. An awareness of the kinds of individual diversity that designers may encounter will thus be helpful to the student of computer science (Shneiderman 1992). Computer professionals need to be able to think about how their decisions will affect the variety of individuals who will use that system, and about how those individuals may in turn influence the use of the technology. Communities and groups. It is rare for a technology to be designed or used by a single individual any more. All communication technologies are designed to connect people, and are usually designed by teams, while the teams themselves are located within organizations. Communities and groups of people have different purposes for their interactions, and there are even disagreements about purpose within those communities and groups. In designing or implementing computer technology the
108
concerns of these groups will influence how the resulting technology is used.
Martin et al.
both how the design is constituted
and
In addition, computer professionals will most likely do the majority of their life’s work within groups or teams. Often they will not be the most powerful person on that team, and will need to make decisions about design and implementation issues within the context of the group interactions and the power hierarchy (Collins and Miller, 1992; Gotterbarn, 1995b; Huff and Jawer, 1994). Both in learning how to work within groups, and in designing technology that will be used by groups, computer professionals need to be prepared to think about social interaction on this level. Organizations. Within any particular organization there are likely to be differing points of view on the use, regulation, promise and design of a technology. These differences will depend on organizational structure and culture, on workplace procedure, on the amount of power and areas of interest each group within the organization has, and on the position in the hierarchy an office or other workplace inhabits. Organizational imperatives that require work groups to value production above all else are often the culprits in poor quality of products or services or even unsafe systems. Every computing professional will either be a member of one or more organization in their career or need to think about technology as it influences or is influenced by these organizations. Cultures. Included under the category of ‘cultural’ divisions are all those groupings one recognizes that cut across local group, organizational and even national boundaries. Thus, gender, race, class and many reference groups, such as ‘hackers’, or ‘computer professionals’ are all examples of cultural groups. These cultural groupings need to be considered by computer professionals when designing and implementing systems. It is also clear that different regional cultures, e.g. European versus Asian, think about ethical issues from different perspectives, and in a marketplace that is increasingly global, designers of systems likely to be used in multiple cultures need to understand the different approaches they take (Swinyard et al., 1990). A recognition of the cultural differences that do exist will help system developers produce and implement systems that are more successful, more widely accepted and perhaps even safer when used in other cultures than that of the designer. Institutional sectors. Different institutions within a culture or a nation will have different interests, preferences, approaches and assumptions as they address particular ethical issues in computing. Governmental agencies, business societies, scientific organizations, charitable or non-profit institutions, professional societies and labour organizations will all differ in their interests and concerns about a particular technology. In addition, any particular technology is likely to be used differently by different institutions, and their interests in regulation of them will be different. Computing professionals will likely be dealing with many of these kinds of institutions in their career, and should be aware that they do not all have similar interests in, or approaches to, technology. Global and national. Social and ethical issues in computing can no longer be considered in isolation from the national and international scene. Given the large scale,
Social and ethical impact of computing
109
global diffusion of technology and the close links between various kinds of technologies in communication networks, even seemingly local decisions about reliability, standards, access, privacy, etc., can have global implications. Many debates about the spread of the Internet assume that ‘democratic values’ will be spread along with the spreading technology. Some countries are quite interested in the technology and connectivity the lnternet represents, but leery of the assumptions of power sharing and democratic process that Western advocates assume are inseparable from the technology (Steele, 1995). Areas of ethical analysis The columns of Fig. 1 represent the variety of ethical issues that confront computer professionals. Several criteria were used to determine what issues to include and how to group them. First, one wanted to make sure that the primary topical issues were included. One did this by making sure to include issues explicitly mentioned in the Computing Curricula 1991 report, and by testing the list of important issues to see if each of them could find a home under one or more of the columns. In addition, one wanted to make sure that each column corresponded to a traditional area of concern for ethicists, and would thus have a rich ethical literature on which to draw. Finally, one attempted to make each column as coherent in its scope as possible, not mixing disparate issues haphazardly or making distinctions too finely. One thinks one has emerged from this process with a set of ethical issues that is coherent, reasonably comprehensive and pedagogically useful. Individual and professional responsibility. It is the authors’ conviction that individual and professional responsibility should be the underpinning to discussions of other specific issues in the table. However, for pedagogical reasons, one felt it useful to include them as two columns set apart for special consideration and foundational to the other issues. It is crucial for each student to examine his/her own conception of individual and professional responsibility, and useful to do it in the context of a course that highlights the social and ethical concerns associated with their chosen profession. Individual responsibilities are those held in common with other people, regardless of technical expertise or position. These include the common moral imperatives listed in the early sections of the ACM code of ethics, such as: avoid harm to others, be honest, be fair, take action not to discriminate, etc. Individual responsibilities are often the result of group memberships, including family, political entities, cultures and professional affiliations. Professional responsibilities are those that computing professionals should undertake because of their special knowledge and skill, their association with others who share that knowledge and skill, and the trust that society places in them because of that knowledge and skill. Computer professionals are not always as easily identifiable as professionals in other occupational groups. To the extent that computing is a profession, its adherents have a responsibility to shape that profession in ways that are at least socially benign, and better, socially desirable. Professional respon-
110
Martin et al.
sibilities, such as those outlined in the ACM code of ethics, are a reasonable list and can be seen to correspond with many of the issues included in the columns of the table. The knowledge of these responsibilities, and the practice of them, is fundamental to ethical thought and behaviour among computer professionals. For this reason, a careful study and application of professional codes of ethics is crucial to ethical practice in computer science. Quality of life. Although few proponents of new technology would introduce their product with a claim that it reduces the quality of life, the concept of quality of life is rarely considered in any detail. In fact, many claims for increases in quality of life are one-dimensional in their assertions, Is faster, better, more, always an increase in quality of life for users of technology? One should consider carefully the possibility that a technology can have some good and some less desirable effects. Under these conditions, the ethical reasoning about the quality of life issues in technology becomes more complex. This kind of reasoning can be difficult, but is necessary to the extent that computer professionals want to be honest about the effects of the technologies they design. Use of power. The knowledge and skill that computer professionals possess gives them power. To the extent that new technology is not totally constrained by physical or mathematical principles, each design decision for that technology is an exercise of power. Computer professionals exercise power in their design of systems and have power exercised over them as members of organizations and work groups. They will become familiar with the ethical dilemmas that face those who are agents of others. They have a responsibility to be a good agent, and a responsibility to those who are affected by their actions. Balancing these responsibilities can produce some of the most excruciating ethical conflicts in the world of computing. Risks and reliability. All technologies are used in a world where consumers, users and the public rely on them to work well. Since error-free design is both impossible to achieve and to measure, computer professionals must become familiar with the inevitable risks associated with technology. Designers of safety critical systems will always have to make choices and trade-offs in design and implementation. Acquaintance with the ethical dimensions of these choices and practice in identifying and evaluating the ethical issues will make for more thoughtful and informed design decisions. Property rights. Even the straightforward issues in the area of property rights are not as simple as they seem. The ACM code of ethics pledges its members to respect copyright and other property laws; but what do we mean by harm in an age when technology makes information infinitely reproducible without degradation of the original? The law on patent, trade secrecy and copyright is mutating to conform to the new technology; and computer professionals who are unaware of the legal and ethical issues are likely to stumble across these issues in even simple design and implementation issues, Careful consideration of both the legal and ethical issues involved in property rights is now essential to the training of computer professionals.
Social and ethical impact of computing
111
Privacy. Computing professionals often design systems that collect, store and transmit data about individuals. Most legal and ethical systems recognize the right to privacy at some level, the right of the individual to determine what data are collected about them by whom and for what purposes. Privacy expectations will differ among individuals, cultures and nations, and need to be taken into account in the design of systems that handle personal data. Some practice in thinking carefully about these issues will help both during design and implementation. Finally, because computing professionals know the potential uses and abuses of the technology, they need to be prepared to participate in the public dialogue about privacy. Equity and access. Careful consideration needs to be given to the extent to which modern technology has divided us into those who have access to the power of technology and those who do not. What responsibilities do computer professionals undertake when they design technology that may dramatically change societal access and power relationships ? Since computer professionals design the systems and influence decisions about implementation, their opinions on these matters should not simply be based on empirical evidence about inequities, but also grounded in careful ethical reasoning about issues of equity and access in society. Honesty and deception. To the extent that computer professionals desire respect and influence in the public area, the respect needs to be based on the public’s expectation of honesty from the profession. Honest, clear and comprehensive appraisals of the benefits and risks of technology will help maintain the respect the public has for the profession. Some reflection on how competing interests balance each other will surely be of help since these issues are likely to confront students early in their careers.
Knowledge units for ethical and social impact of computing A knowledge unit defines a coherent collection of subject matter that is so fundamental to the designated subject area that it should occur in every undergraduate curriculum. Frequently, the subject matter of a particular knowledge unit is related to other knowledge units in the common requirements and can be introduced within any of several alternative course structures. This is often the case with the knowledge units that one described in this new subject area designated Ethical and Social Impact of Computing (ES). The conceptual framework shown in Fig. 1 is used as the basis for the knowledge space of this new subject area. The five fundamental will be defined as: l l l l l
ESl: ES2 ES3: ES4: ES5:
knowledge
responsibility basic elements basic skills of basic elements basic skills of
units of Ethical and Social Impact
of the computer professional, of ethical analysis, ethical analysis, of social analysis, and social analysis.
of Computing
112
Martin et al.
It is important to note that these knowledge units should be taught in a computing context using topics related to computing, such as privacy, automation, freedom of speech on networks, risks and reliability, and equity of access to computer technology. ESl: responsibility of the computer professional Personal and professional responsibility must be the foundation for discussions of all topics in this subject area. Since student response to specific issues will be at least partly determined by the ways they understand their own individual responsibility, it is crucial for each student to understand that the ethical principles of honesty, fairness, autonomy, justice and beneficence define personal responsibility. Personal responsibilities are those held in common with other people, regardless of technical expertise or position. They should not be thought of as the obligations of socially isolated individuals, for they are often the result of group memberships, including family, political entities, cultures and professional affiliations. Professional responsibilities are those additional responsibilities that computing professionals should undertake because of their special knowledge and skill, their association with others who share that knowledge and skill, and the trust that society places in them because of that knowledge and skill. The knowledge of these responsibilities, and the practice of them, is fundamental to ethical thought and behaviour among computer professionals. The five areas to be covered under the responsibility of the computer professional are: (1) (2) (3) (4) (5)
history of the development and impact of computer technology, why be ethical? major ethical models, definition of computing as a profession, and codes of ethics and professional responsibility for computer professionals.
History of the development and impact of computer technology. To set the stage for an understanding of the professionalism, ethics and social impact of computing, it is necessary for students to see how computing has evolved in the historical and social context. Tracing the history of the mechanization of computation, the development of programmable devices, and the evolution of information representation, transmission and storage will help students to understand how computers are a cultural artefact with profound social impact. When students realize how young the computing profession is relative to other professions and how rapidly it is changing, they have a better appreciation about why there are so many unresolved ethical and social issues with which they will have to deal. Why be ethical? It is necessary to make a strong case for ethical behaviour in the context of professional practice. Many computer science students enter college with a hacker mentality, so it is important to help them to become aware of the tremendous responsibility to other people that comes with the practice of their
Social and ethical impact of computing
expertise. Analogies with medicine, engineering stand the importance of ethical behaviour.
113
and the law help students to under-
Major ethical models. In order to understand the basis for personal and professional responsibility, students need to learn about several major ethical models that can be used to evaluate alternatives and decision-making. Ethical models developed by philosophers, such as Bentham’s Utilitarianism, Kant’s Moral Imperative, and Rawles’ negotiation of social contracts, have been used effectively to provide a formal basis for discussion of ethical dilemmas. Definition of computing as a profession. A major criterion for ‘professional’ status is the expectation of some autonomy in the exercise of responsibilities. Some computer professionals have a great deal of autonomy in the decisions they make. Other computer professionals are employees of large firms and software shops where they are quite restricted in their duties and autonomy. They often work on only a small part of a larger system and have little knowledge of or control over decisions made about the larger project. The extent to which computer professionals are ‘professional’ is more analogous to accountants or engineers who claim special expertise, have professional associations and adhere to a code of ethics and professional conduct. As they prepare for a computing career, it is very important for computing students to realize early in their education that they will be entering a profession, not just a job market. Codes of ethics and professional responsibility for computer professionals. To the extent that computing is a profession, its adherents have a responsibility to shape that profession in ways that are of benefit to society. An important part of the evolution of computing into a profession has been the development of codes of ethics and professional practice which delineate the responsibilities of the computer professional. The knowledge of these responsibilities, and the practice of them, is fundamental to ethical thought and behaviour among computer professionals. For this reason, a careful study and application of professional codes of ethics is crucial to ethical practice in computer science. ES2: basic elements of ethical analysis Three basic elements of ethical analysis that students to use in their decision-making are:
need to learn and be able
(1) ethical claims can and should be discussed rationally, (2) ethical choices cannot be avoided, and (3) some easy ethical approaches are questionable. Since students will need to defend their opinions and decisions when they work within organizations and teams, they need to be able to discern the reasons behind their opinion and decisions and to evaluate them. Ethical analysis should be introduced in a module dedicated to the presentation of classical ethical frameworks and principles. The ethical dimensions of design and implementation decisions should be discussed in technical classes so that students will learn to view ethical
114
Martin et al.
decision-making as a crucial part of system development. Students will see that, in some cases, it is easy to determine correct behaviour for a computer professional; and, in other cases, a careful ethical analysis is needed to identify best choices in situations that raise difficult ethical issues. Ethical claims can and should be discussed rationally. Since many individuals have strong emotional responses or moral intuitions about the ethical issues confronting computer professionals, discussions on these issues can quickly disintegrate into entrenched opinions or even hostility unless students realize that ethical claims can and should be discussed rationally. Even the most deeply felt emotional responses can, with careful analysis, be articulated in terms of their structure and supporting reasons. Reasons can also be evaluated in terms of criteria like consistency, logical coherence, agreement with accepted standards (e.g. codes of ethics), and applicability to a variety of cases. None of these criteria are infallible methods of evaluation, but they are rational. Since students will need to defend their opinions and decisions when they work within organizations and teams, they need to be able to discern the reasons behind their opinion and decisions and to evaluate them. Evaluation of ethical arguments helps one to be reasoned in the position taken, rather than simply letting personal preference, convenience, bias, or apathy determine an ethical position. Ethical choices cannot be avoided. Many students think that ethical choices are irrelevant to a computer science or engineering discipline, and that it is their job simply to do their assigned work tasks. Careful ethical thought involves recognizing that moral responsibility is an important aspect of computer technology and cannot be abandoned by choice. Courts of law, the military code and many professional societies’ codes do not allow voluntary moral retreat, and computing professionals should be aware of its dangers. One enters the realm of ethical choice in system development any time one makes decisions that affect people. Ethical reflection should begin with the assumption that all design and implementation involves value choices. Some easy ethical approaches are questionable. Students need to become aware of the differing grounds for ethical claims that have become common, and of the significant weaknesses of those often put forward in defence of an ethical choice. There are a variety of naive approaches to ethical reasoning that may lead the beginning ethical thinker astray: (1) Naive relativism: the belief that all moral choices are relative to the situation and the culture makes it easy to get along in polite conversation with others, since it requires no confrontations. However, when difficult choices have to be made, students need to realize that a truth can emerge that is not culturally specific. (2) Naive egoism: the simple belief that selfishness is the best guiding principle can make it convenient to ignore duty to others while concentrating only on personal profit. This approach conceals a fundamental inconsistency, since its naive form suggests that everyone else should still follow ethical forms.
11.5
Social and ethical impact of computing
(3) Naive agency: surrendering all moral authority by claiming to be a simple agent of some other entity, e.g. your employer, has its own clear problems. In the end, even the legal system requires individual responsibility, and military codes of conduct require soldiers to disobey some orders. Responsibility cannot be this easily dismissed. (4) Naive legalism: equating ethicality with legality is a tempting way to dispense with serious ethical reflection. Students should certainly be aware of the legal issues that will confront them. However, assuming that ‘if it is legal, it is ethical’ is asking more of the law than it can provide, and denies the legitimacy of principled disagreement with the law. ES3: basis skills of ethical analysis Five basic skills of ethical analysis that will help the computer apply ethics in their technical work are: (1) (2) (3) (4) (5)
science student
to
arguing from example, analogy and counter-example, identifying stakeholders in concrete situations, identifying ethical issues in concrete situations, applying ethical codes to concrete situations, and identifying and evaluating alternative courses of action.
Since these are basic tools of ethical reasoning, computing professionals need practice in using them in relation to technical issues. These skills should be introduced in an ethics module early in the curriculum and students should be given enough time to move from their own intuition, to specific reasons, to concrete cases and back again in intense discussions with other students and faculty. They should then be used in subsequent technical courses with case studies and design exercises. Arguing from example, analogy and counter-example. Moral argument often takes place using examples, counter-examples and analogies. These methods help to clarify issues and to point out the inconsistencies and difficulties inherent in various ethical positions. Ethical argument moves from intuition about the right or the good, to explicit reasoning, and then is tested by being compared to concrete examples and analogies. Identifying stakeholders in concrete situations. When considering a particular implementation of a technology, a common mistake is to accept a simple definition of who the relevant affected parties are. It is only afterwards, and too late, that one realizes some important parties were not considered or consulted. Students need practice in being able to identify all the stakeholders in a given context as early as possible in the system development process. Identifying ethical issues in concrete situations. Another common pitfall is to stop reasoning about a particular decision when one has found one good reason to act or to refrain from action, usually due to the pressure of time and deadlines. Only
116
Martin et al.
in hindsight does one determine that other important ethical issues were also at stake. Practice in identifying ethical issues in a situation will train students to have more foresight as practising computer professionals. Applying ethical codes to concrete situations. Professional ethical codes represent the consensus of the field with regard to the responsibilities of those professionals. Students who aspire to be part of the computing profession should be familiar with ethics codes in their application to situations that are likely to occur in professional life. The methods of ethical reflection one is asking students to learn include comparing their intuitions and reasoning to both concrete cases and to established values in the field. Ethical analysis that does not do this is incomplete, because it is not anchored in dialogue with the ethical reasoning of other professionals. It should be clear from the way one is proposing dialogue, that one is talking about thoughtful reflection, with significant concern for the consensus of the field on ethical issues. Identifying and evaluating possible courses of action. Often, ethical mistakes are made more from a failure of imagination rather than a failure of moral principle. If you simply cannot think of alternative courses of action, then you are left with little choice. An important practical aspect of ethical reasoning involves determining if all the available options have been exhausted. Many students will look at a situation and assume that a few obvious and often dichotomous choices are the only choices. Practice in recognizing a range of options can give students insight to think about choices. Practice in evaluating options they have generated can provide students with the confidence they will need to negotiate solutions in actual situations.
ES4: basic elements of social analysis Five basic elements of social analysis are: (1) (2) (3) (4) (5)
the social context influences the development and use of technology, power relations are central in all social interaction, technology embodies the values of the developers, populations are always diverse, and empirical data are crucial to the design and development processes.
Students need to be presented with examples and case studies that illustrate impact of social issues in system development.
the
The social context influences the development and use of technology. Technology does not simply ‘impact’ society in a one-way causal chain, but society also influences the shape and development of technology. In addition, the social or organizational setting in which a technology is used influences the way it is used. Students need to be aware that anticipating the use of computer systems in context is an important part of the development process.
Social and ethical impact of computing
117
Power relations are central in all social interaction. All social relationships have implicit and explicit considerations of power. Developers of technology need to be aware of the relative power of different parties in an organization as they develop systems for that organization. They also need to be aware of how those power relationships may shift as a result of the new technology. Technology embodies the values of the developers. Technology transmits and embodies assumptions and values when choices are made during the development stage. For example, decisions to adopt particular standards, to use particular methods, to implement particular features, to adhere to particular criteria, and even to design to the specification are all ethical decisions. It is unethical to ignore the values embedded in these technical artefacts. Populations are always diverse. The situations in which a technology will be used, the people who will use that technology, and the uses to which it will be put, are all more varied and diverse than one might first expect. It is better to develop the system with this variability in mind than to be surprised by it later. Students need to learn to consider issues such as culture, gender, ethnicity, age and disability at the very beginning of every design process. Empirical data are crucial to the design and development processes. Models never capture reality completely: a computer professional must systematically collect and analyse data to guide the development process. Speculation, hunches or naive optimism about a particular social context are not acceptable substitutes for empirical data. ES.5 basic skills of social analysis Three basic skills of social analysis appropriate (1) identifying tion, (2) identifying (3) evaluating, nology.
and interpreting
for computer
the social context
professionals
of a particular
are:
implementa-
assumptions and values embedded in a particular system, and by use of empirical data, a particular implementation of a tech-
Identifying and interpreting the social context of a particular implementation. Quality system development must take into account the people and organizations that will interact with the system. This information is typically not described fully in system specifications. Students should be provided with opportunities to analyse the social context surrounding actual computer implementations. Identifying assumptions and values embedded in a particular system. The claim that a particular implementation of a technology is value-neutral is typically based on a failure of imagination. Practice in identifying embedded values will help students to identify similar issues in the future.
118
Martin et al.
Using empirical data to evaluate a particular implementation of technology. Practising computer professionals need to be able to use empirical data to evaluate the likely use of a technology (as opposed to its planned use) and the performance of a technology after its implementation. The designer of a technology needs enough data to help determine if (within constraints of time and budget) the design has the effects in actual use it claims to and has no other significant risks associated with it. Students need to be provided with opportunities to practice doing evaluations and to be shown how to make professional judgement calls about how much and what quality of evaluation to do.
Using knowledge units to develop a curriculum The strategy chosen by a particular programme to implement this new subject area should be pedagogically driven. It should fit into the rest of the programme in an integrated fashion so that the relationship between the knowledge units in this area and the rest of the curriculum is apparent to students. The five knowledge units above can become part of a computer science curriculum in many different ways. In this section three ways are described: (1) the five knowledge units in a course dedicated to computer ethics and social impact (Friedman and Winograd, 1990; Martin and Holz, 1991) are included; (2) individual topics and activities from the knowledge units in existing courses, which integrate the computer ethics by focusing on examples and motivation from the subject areas important to the course (Miller, 1988) are included; and (3) several of the knowledge units are amalgamated into a ‘capstone course’, a senior level project course that emphasizes skills and knowledge required for becoming a responsible computer professional (Cotterbarn, 1991). This list of three strategies is neither exhaustive, nor mutually exclusive. However, it does reflect three strategies described in computer science education literature. A course dedicated to computer ethics and social impact: some schools have a course for either computer science majors or for the general student population, dedicated to computer ethics or computers and society, or technology issues. Such a course could comfortably include the lectures and laboratory studies described in the five knowledge units. More hours could also be added that included in depth study of several case studies or emerging technologies and their ethical significance. Integrating ethics and social impact into existing courses: a faculty member can include topics and activities from the ES knowledge units into the more traditional technical classes. A lecture about testing can be enlivened by a discussion about human costs (that are often ignored by developers), while deciding how long to continue testing. Ethical issues can be described in conjunction with a programming assignment. For example, students could be asked to prioritize a list of organ transplant candidates. The students would have to program and justify a priority scheme of their own choosing.
Social and ethical impact of computing
119
Ethics and social impact in a capstone course: a project driven course is an ideal place to include professional and ethical issues. The course can be a senior practice, software engineering project course, or an ethics course directed at seniors. The ethical issues arise from two sources in these courses: either within the various responsibilities at each stage of the software development life cycle, or from more general professional obligations. The general structure of such a course is to have the class form project teams to work on large projects. The sense of professional obligation to quality products and to fellow software developers can be emphasized. These issues are important to industry. Another strategy is to have students develop a Social Impact Statement, similar to an Environmental Impact Statement, in which students think and write about the social context in which the project is (or might be) embedded and to evaluate how that context effects the performance of the product. It is important for the ES knowledge units to be covered more than once, or in more than one place in the curriculum, if they are to be really understood by students. The power of the conceptual framework one provides does not stem from its restrictiveness (it is remarkably flexible), but from the grounding that it provides for a variety of implementations within the constraints of a standard computer science undergraduate programme. The importance of laboratory work: computer science is fundamentally a laboratory orientated discipline. Most courses incorporate laboratory work in one form or another, e.g. ‘open labs’ with programming assignments done outside of class or ‘closed labs’ with assignments done in class. Since the emphasis in ethical and social issues in computing is on computers in the context of their use, one encourages instructors to incorporate laboratory work in courses in this area. This can be done in two ways. First, other courses in database design, human computer interaction, operating systems, algorithms, etc., can incorporate ethical and social issues into their laboratory work. There are several readily available guides to how this might be done (Gotterbarn, 1995b). Many other knowledge units in computer science are woven into the curriculum in this manner, e.g. database principles are sown throughout the curriculum, and gain their strength from this repeated and varied emphasis. Ethical and social issues can and should be scattered throughout the laboratory curriculum in a similar manner.
The minima! implementation A practical question at this point is ‘what is a minimally sufficient implementation of the knowledge units ?’ This question has several layers, and these are addressed separately before attempting a general answer. First, any sufficient implementation of this subject area will cover all the social and ethical basic elements and skills, at least at the introductory level. Leaving out any significant number of them would drastically reduce the coherence of this curriculum strand. One also believes strongly that an understanding of personal and professional responsibility is fundamental to ethical analysis. Secondly, one suggests that a sufficient implementation should include several of the ethical issues and several of the levels of social
Martin et al.
120 Table
I. Minimal implementation
of ES knowledge units
Knowledge units
Lecture hours
Laboratory hours
ES1 ES2 ES3 ES4 ES5 In-depth topics
3 3 3 3 3 10
4 3 5 3 5 5
Total hours
25
25
analysis. This is suggested partly because an understanding of any one of them will inevitably lead to consideration of the others, and one cannot conceive of a course that dealt with one ethical issue with no mention of or recourse to other ethical issues or to specific levels of social analysis (Ermann et al., 1990; Kling and Dunlop, 1991; Rosenberg, 1992; Huff and Finholt, 1994; Johnson and Nissenbaum, 1995). The amount of time spent dealing with the knowledge units is important: a minimum of 15 lecture hours and 20 laboratory hours of the curriculum should be devoted to this material in order for students to gain an in-depth understanding of the basic elements and skills. In addition, it is strongly recommended that another 10-15 lecture hours be spent on an in-depth coverage of topics such as privacy, computers in medicine, deskilling of jobs and computer crime to enable the students to apply the basic elements and skills to real issues (see Table l). CONCLUSIONS One believes that this report provides a useful and coherent conceptual framework for the presentation of ethical and social issues in computing to students of computer science. It provides instructors and students with a context for considering these issues. It integrates both the literature on the impact of computing on society and the ethics literature into a package that helps each approach illuminate the other while providing balance and perspective. Finally, it provides a clear set of learning objectives in the form of knowledge units and an organized menu of topics for instructors to use. They provide a foundation for further thinking and practice in the undergraduate, graduate and business world. The approach to ethical and social issues in computing presented in this report can invigorate the teaching of these topics at the undergraduate level. One is committed to both integration of topics and a specialized course, to the consideration of several ethical issues and levels of social analysis, and most importantly, to the full inclusion of the principles and skills. It is the vision of the ImpactCS Steering Committee that the approach to integrating ethical and social impact issues into computer science proposed by this report will make a positive and profound change to the teaching and practice of computer science as one enters the 21st century.
Socialand ethical impact of computing
121
Project ImpactCS steering committee Ron Anderson, University of Minnesota; John Artz, George Washington University; David Bellin, North Carolina A & T State University; Tim Bergin, American University; Louis Berzai, University of Notre Dame; Terre11 Ward Bynum, Research Center on Computing and Society; Peter Caws, George Washington University; Frank W. Connolly, American University; Batya Friedman, Colby College; Donald Gotterbarn, East Tennessee State University; Charles Huff, St Olaf College; Deborah Johnson, Rensselaer Polytechnic Institute; Rob Kling, University of California, Irvine; Blaise Liffick, Millersville University; Joyce Currie Little, Towson State University; Walter Maner, Bowling Green State University; C. Dianne Martin, chair, George Washington University; Keith Miller, University of Illinois, Springfield; Michael Mulder, University of Southwestern Louisiana; Eric Roberts, Stanford University; Richard S. Rosenberg, University of British Columbia; Ronni Rosenberg, Consultant; Patrick Sullivan, Computer Ethics Institute; Albert (Joe) J. Turner, Clemson University; Harry Yeide, George Washington University. ACKNOWLEDGEMENTS
The work of Project ImpactCS has been supported by the National Science Foundation grant DUE-9 354 626 and by The George Washington University. We would also like to thank the outside readers for their insightful comments on the first ImpactCS report. They are: Jacques Berleur, Loyola University, Chicago; Dorothy Denning, Georgetown University; Charles E. M. Dunlop, University of Michigan, Flint; Joseph Herkert, North Carolina State University; and Margaret Anne Pierce, Georgia Southern University. REFERENCES ACM/IEEE-CS Joint Curriculum Task Force (1991) Computing Curricula 1991. ACM/IEEE-CS Joint Curriculum Task Force Report. New York: ACM Press. ACM Curriculum Committee on Computer Science (1979) Curriculum 78: Recommendations for the undergraduate program in computer science. Communications of the ACM 22(3), 147-66.
Collins, R. and Miller, K. (1992) Paramedic ethics for computer professionals. Journal of Systems and Software January 1992, l-20. Computer Sciences Accreditation Board (1987) Criteria for accrediting programs in computer science in the United States. Technical report. Ermann, M. D., Williams, M. B. and Gutierrez, C. (1990) Computers, Ethics, and Society. New York: Oxford University Press. Friedman, B. and Winograd, T. (eds) (1990) Computing and Social Responsibility: A Collection of Course Syllabi. Palo Alto, CA: Computer Professionals for Social Responsibility. Gotterbarn, D. (1991) The capstone course in computer ethics. In Proceedings of the National Conference on Computing and Values, New Haven, CT. Gotterbarn, D. (1995a) Responsibility regained. In D. G. Johnson and H. Nissenbaum (eds) Computer Ethics and Social Values. pp. 18-24. Englewood Cliffs, NJ: Prentice Hall. Cotterbarn, D. (1995b) A tool kit of computer ethics activities for computer science classes. In Proceedings of National Educational Computing Conference, Baltimore, MD.
122
Martin et al.
Huff, C. W. and Finholt, T. (1994) Social Issues in Computing: Putting Computing in its Place. New York: McGraw-Hill. Huff, C.W., and Jawer, B. (1994) Toward a design ethic for computing professionals. In C.W. Huff and T. Finholt (eds) Social Issues in Computing, pp. 691-713 New York: McGrawHill. ImpactCS Steering Committee (May, 1995) Consequences of Computing: A Framework for Teaching the Social and Ethical Impact of Computing. First Report of the Project ImpactCS Steering Committee. GWU-ICTSP Technical Report No. 95-03, George Washington University, Washington, DC. ImpactCS Steering Committee (1996) Implementing the Tenth Strand: Extending the Curriculum Requirements for Computer Science. Second Report of the ImpactCS Steering Committee. GWU-ICTST Technical Report No. 96-01, George Washington University, Washington, DC 20052. Johnson, D. G. and Nissenbaum, H. (eds) (1995) Computer Ethics and Social Values. Englewood Cliffs, NJ: Prentice Hall. Kling, R. and Dunlop, C. (1991) Computerization and Controversy: Value Conflicts and Social Choices. New York: Academic Press. Martin, C. D. and Holz, H. J. (1991) Non-apologetic computer ethics education: A strategy for integrating social impact and ethics into the CS curriculum. In Proceedings of the National Conference on Computing and Values, New Haven, CT, 50-66. Miller, K. (1988) Computer ethics in the curriculum. Computer Science Education 1, 37-52. Rosenberg, R. S. (1992) The Social Impact of Computers. New York: Academic Press. Shneiderman, B. (1990) Human values and the future of technology: A declaration of empowerment. Computers & Society 20(3),1-6. Shneiderman, B. (1992) Designing the user interface. Reading, MA: Addison-Wesley. Steele, F. R.(1995) Planning a global human cyberspace. The Second NCLS Conference on Legal, Ethical, and Technological Aspects of Computer Use and Abuse. Conference Report, Washington, DC: AAAS. Swinyard, W. R., Rinne, H. and Keng Kau, A. (1990) The morality of software piracy: a crosscultural analysis. Journal of Business Ethics, 9(8), 655-64.
