Font Size:
Ethics in the Internet Design Process
Last modified: 2011-03-25
Abstract
Within a few months of the first US government grant to develop networked computing, in 1969, a technical document series was launched to record the decision-making processes of those who were designing what we now know as the Internet. The Internet Requests for Comments (RFCs) series continues today, including well over 6,000 items by 2011.
The computer scientist and electrical engineer authors of Internet RFCs thoroughly understood that "Network topology is a complicated political and economic question . . . " (RFC 613, p. 1) involving important ethical problems. This paper presents an analysis of the ways in which those responsible for technical design of the Internet confronted ethical issues by addressing the following questions:
* Which technical problems in the design of the Internet were perceived to have ethical dimensions?
* What ethical issues were deemed important enough to require design-based solutions?
* Under what conditions did ethical issues become apparent to those responsible for technical design of the Internet?
* How do Internet designers define such concepts as "ethical," "normative," "fair," and "just?"
* How did perceived differences in types of users and uses affect ethical decision- making?
* How did designers of the Internet conceptualize and operationalize ethics for non- human users of the network?
The Context
Developments both internal to the RFC discourse and external to it provided the context within which ethical issues were addressed. Policy frames developed within the design discourse offered a conceptual foundation for addressing concrete ethical dilemmas. Social trends and specific events heightened sensitivity to ethical issues as time went on.
Policy frames. In the course of solving technical problems, those involved in creating the Internet discussed a number of legal and policy issues, engaged in explicit and implicit policy analysis and policy-making, developed formalized decision-making processes and entities, and put forth political ideas. Policy frames were developed for thinking through such matters in three areas.
(1) Users: Policy frames regarding users were evident in the areas of attitudes toward users, distinctions among types of users, and support for users. Distinctions among types of users with relevance for analysis of ethical issues included benign vs. malicious users, insiders vs. outsiders, and – the most important of all – human vs. non-human (daemon) users. The latter category includes hardware processes, software, and operational layers of the network.
(2) Uses: A wide range of both governmental and non-governmental uses was taken into account from the start by network designers, who were well aware that ultimately the Internet would be commercialized.
(3) Design criteria: Design criteria for the Internet design process can be considered policy principles. Logistical criteria include content reliability, network reliability, and multidirectional compatibility. Social design criteria include network proprioception, technological democracy, telepresent distant and distributed computing, and stimulating innovation qua innovation. Ethical issues are confronted during the operationalization of both types of design criteria. Queuing rules for network traffic were required, for example, in order to meet the criterion of network reliability, but it was well understood that the socio-technical decision-making required to achieve this had significant ethical implications as well.
An insistence on technological democracy was a social design criterion of particular importance from an ethical perspective. Those responsible for technical design of the Internet were working on the technological horizon, developing computers and networks with the greatest capacity and speed. At the same time, though, they recognized that it was imperative to ensure that the network could accommodate those at the opposite end of the spectrum of technological advantage, with limited computing and network capacity. It was this insistence on simultaneously serving the cutting edge and those with the fewest resources that made it possible for the Internet to be taken up as quickly as it was around the world.
Stimuli for ethical sensitivity. A variety of factors combined to enhance ethical sensitivity over time. Within computer science and electrical engineering, the need to adhere to professional standards of behavior had long been acknowledged but did not generate much activity until the early 1980s. In 1981, computer scientists concerned about their role in the defense industry during a period of concern over a growing nuclear threat came together to discuss ethical issues in computing, creating a group that ultimately becaame Computer Professionals for Social Responsibility (CPSR).
Within the first decade of the Internet design process (1969-1979), there was no explicit discussion of issues labeled as ethical although numerous policy issues received attention. It took the first Internet-wide experience with a destructive worm, in November of 1988, to really bring ethical concerns into the design conversation. This experience was so traumatic that it quickly triggered the development of codes of ethics by, among others, an Internet governance organization, the Internet Advisory Board (IAB); the US government's National Science Foundation (NSF) as the funder of Internet development; MIT as an academic institution heavily invested in the development and use of the Internet; and CPSR.
The US Federal Bureau of Investigation (FBI) takedown and ultimate destruction of a small computer company in 1990 in pursuit of what turned out to be completely legal activity raised concern about the public interest in networked computing among non-specialists and led to the formation of the influential non-profit organization, the Electronic Frontier Foundation (EFF).
Commercialization of the network, in 1993, was another turning point; designers understood even before the fact that bringing so many new types of users and uses into the network would inevitably also introduce many new ethical problems.
Research Method
The analysis presented in this paper is based on an inductive reading of several hundred documents in the RFC series identified through a search of the entire corpus using the terms "ethics," "ethical," "norm," "normative," "fair," "fairness," "just," "justice," "moral," "equity," "good" and "bad" behavior, and "good citizenship." Theories and concepts from three literatures -- philosophy, communication, and science, technology, and society -- are used in the analysis.
The study is part of a larger project funded by the NSF involving a comprehensive inductive reading of the entire corpus of the Internet RFCs analyzing perceptions of and positions on policy issues, the use of political and social theory by those responsible for technical design of the Internet, and the development of formal policy-making processes and institutions for global Internet governance through what began as an informal conversation among graduate students.
Findings
Analysis of the pertinent items in the RFCs up to the point of commercialization, through the close of 1992, provides insight into the variety of ethical concerns that the computer scientists and electrical engineers believed needed to be addressed in the course of the design process and justifications for asserting ethical standards. A variety of mechanisms that could be used to ensure that the network functioned in an equitable manner were identified, but these technical thinkers were also aware that, ultimately, human behavior and the law would be pertinent.
Ethical concerns. The first ethical issues raised involved what participation in the network would require of individual "hosts," or users. As a matter of equity, designers sought to create technical standards ("protocols") that would neither demand significant changes in existing systems nor create a situation in which every given instantiation of a technology would have to incorporate all possible protocols.
The bulk of discussion about ethical concerns, however, involved resource use. It was believed that all connections should be treated in the same way. A problem in one among the "network of networks" comprising the system shouldn't be allowed to bring other networks down. Industrial research organizations were expected to pay their own way, to ensure that commercial concerns didn't degrade the quality of networking for those government agencies and research institutions for whom the Internet was primarily intended during its early years. Those doing research on the network itself were admonished to make sure that their experimentation did not itself cause damage to the network. Access issues, such as pricing policies, were also discussed in ethical terms.
Justifications for ethical standards. The philosophical question of the differences between concepts such as "MUST," "SHALL," and "CAN" is foundational. Use of such terms in programming interacted with discussions about just how compliance with technical standards might be achieved and with the growing formalization of the legal status of decision-making for the Internet. Ultimately the concept of "normative" came to mean that compliance with a given protocol was required, and RFC documents presenting official standards included a section on their normative contexts.
As early as 1989, the argument that the network is a type of infrastructure as important to society and the economy as the road system was used as a justification for insisting on ethical standards in its design and use. The fact that US government funds were essential to its development also meant that adherence to NSF rules regarding unethical and unacceptable activity could be required. Many believed the network should be treated as a common carrier, subject to the regulatory mandates to treat all users in the same manner and not to distort the content in any way. The internationalization of the network, which began in the early 1970s, brought intercultural dimensions of ethical questions into play.
Most of these arguments for insisting upon ethics in Internet design and use remain valid today, and additional political, social, cultural, and economic arguments came into play once the network was commercialized. One justification used in early years is no longer valid, however. In 1991 it was possible to take the position that since no one had to be on the Internet, it was reasonable to require adherence to ethical standards as a quid pro quo for the privilege of being online, but this argument cannot be made a decade into the 21st century.
Mechanisms for ensuring network fairness. The bulk of the techniques for incorporating ethical standards into the network discussed in the RFCs were, of course, technical. These appeared at multiple layers of the network, including at the point of linkage to the network, at gateways between networks, at packet switches. The single technical problem involving equity that received the most treatment up to the point of commercialization was "fair queuing," the issue of how to make sure that the sequencing of messages through the network took place in an equitable manner. Several different approaches were explored. There was concern about whether or not it would be possible to absolutely ensure equity in network use given differences in types of content, uses, and the distance messages have to travel.
Interactions between technology, law, and society. Social policy and behavioral expectations were also treated as mechanisms for ensuring network fairness. The right to have access to encryption, for example, and pricing policies were considered. Specific guidelines were developed to ensure that research on the network wuold not generate ethical problems in the course of experimentation.
Internet designers were aware that it would be possible for bad actors to "game" a system intended to ensure equity. A distinction was drawn between learning how to use a computer system and learning how to use it properly, with the latter including making sure that one is not offending other users. Ethics and etiquette became linked. Computer crime was understood to involve an intertwining of social, legal, and technical matters.
Notions of "good citizenship" came to inform both design decisions and behavioral expectations. The good network citizen shares resources, even when doing so comes at a cost to him- or herself; respects the content, communication, and data of others; respects the resources of others; and complies with protocols and programming rules.
Contributions
This research offers insight into the ways in which specific ethical issues became visible to technical decision-makers as mandates affecting Internet design. It enriches our understanding of the human side of the design process, itself now a model for those building other large-scale socio-technical systems. It is hoped that uses of these findings will help bring technical decision-makers into a common conversation with those in the social sciences and humanities. Finally, the project furthers the development of research methods for analysis of technological discourses with ethical consequences.
The computer scientist and electrical engineer authors of Internet RFCs thoroughly understood that "Network topology is a complicated political and economic question . . . " (RFC 613, p. 1) involving important ethical problems. This paper presents an analysis of the ways in which those responsible for technical design of the Internet confronted ethical issues by addressing the following questions:
* Which technical problems in the design of the Internet were perceived to have ethical dimensions?
* What ethical issues were deemed important enough to require design-based solutions?
* Under what conditions did ethical issues become apparent to those responsible for technical design of the Internet?
* How do Internet designers define such concepts as "ethical," "normative," "fair," and "just?"
* How did perceived differences in types of users and uses affect ethical decision- making?
* How did designers of the Internet conceptualize and operationalize ethics for non- human users of the network?
The Context
Developments both internal to the RFC discourse and external to it provided the context within which ethical issues were addressed. Policy frames developed within the design discourse offered a conceptual foundation for addressing concrete ethical dilemmas. Social trends and specific events heightened sensitivity to ethical issues as time went on.
Policy frames. In the course of solving technical problems, those involved in creating the Internet discussed a number of legal and policy issues, engaged in explicit and implicit policy analysis and policy-making, developed formalized decision-making processes and entities, and put forth political ideas. Policy frames were developed for thinking through such matters in three areas.
(1) Users: Policy frames regarding users were evident in the areas of attitudes toward users, distinctions among types of users, and support for users. Distinctions among types of users with relevance for analysis of ethical issues included benign vs. malicious users, insiders vs. outsiders, and – the most important of all – human vs. non-human (daemon) users. The latter category includes hardware processes, software, and operational layers of the network.
(2) Uses: A wide range of both governmental and non-governmental uses was taken into account from the start by network designers, who were well aware that ultimately the Internet would be commercialized.
(3) Design criteria: Design criteria for the Internet design process can be considered policy principles. Logistical criteria include content reliability, network reliability, and multidirectional compatibility. Social design criteria include network proprioception, technological democracy, telepresent distant and distributed computing, and stimulating innovation qua innovation. Ethical issues are confronted during the operationalization of both types of design criteria. Queuing rules for network traffic were required, for example, in order to meet the criterion of network reliability, but it was well understood that the socio-technical decision-making required to achieve this had significant ethical implications as well.
An insistence on technological democracy was a social design criterion of particular importance from an ethical perspective. Those responsible for technical design of the Internet were working on the technological horizon, developing computers and networks with the greatest capacity and speed. At the same time, though, they recognized that it was imperative to ensure that the network could accommodate those at the opposite end of the spectrum of technological advantage, with limited computing and network capacity. It was this insistence on simultaneously serving the cutting edge and those with the fewest resources that made it possible for the Internet to be taken up as quickly as it was around the world.
Stimuli for ethical sensitivity. A variety of factors combined to enhance ethical sensitivity over time. Within computer science and electrical engineering, the need to adhere to professional standards of behavior had long been acknowledged but did not generate much activity until the early 1980s. In 1981, computer scientists concerned about their role in the defense industry during a period of concern over a growing nuclear threat came together to discuss ethical issues in computing, creating a group that ultimately becaame Computer Professionals for Social Responsibility (CPSR).
Within the first decade of the Internet design process (1969-1979), there was no explicit discussion of issues labeled as ethical although numerous policy issues received attention. It took the first Internet-wide experience with a destructive worm, in November of 1988, to really bring ethical concerns into the design conversation. This experience was so traumatic that it quickly triggered the development of codes of ethics by, among others, an Internet governance organization, the Internet Advisory Board (IAB); the US government's National Science Foundation (NSF) as the funder of Internet development; MIT as an academic institution heavily invested in the development and use of the Internet; and CPSR.
The US Federal Bureau of Investigation (FBI) takedown and ultimate destruction of a small computer company in 1990 in pursuit of what turned out to be completely legal activity raised concern about the public interest in networked computing among non-specialists and led to the formation of the influential non-profit organization, the Electronic Frontier Foundation (EFF).
Commercialization of the network, in 1993, was another turning point; designers understood even before the fact that bringing so many new types of users and uses into the network would inevitably also introduce many new ethical problems.
Research Method
The analysis presented in this paper is based on an inductive reading of several hundred documents in the RFC series identified through a search of the entire corpus using the terms "ethics," "ethical," "norm," "normative," "fair," "fairness," "just," "justice," "moral," "equity," "good" and "bad" behavior, and "good citizenship." Theories and concepts from three literatures -- philosophy, communication, and science, technology, and society -- are used in the analysis.
The study is part of a larger project funded by the NSF involving a comprehensive inductive reading of the entire corpus of the Internet RFCs analyzing perceptions of and positions on policy issues, the use of political and social theory by those responsible for technical design of the Internet, and the development of formal policy-making processes and institutions for global Internet governance through what began as an informal conversation among graduate students.
Findings
Analysis of the pertinent items in the RFCs up to the point of commercialization, through the close of 1992, provides insight into the variety of ethical concerns that the computer scientists and electrical engineers believed needed to be addressed in the course of the design process and justifications for asserting ethical standards. A variety of mechanisms that could be used to ensure that the network functioned in an equitable manner were identified, but these technical thinkers were also aware that, ultimately, human behavior and the law would be pertinent.
Ethical concerns. The first ethical issues raised involved what participation in the network would require of individual "hosts," or users. As a matter of equity, designers sought to create technical standards ("protocols") that would neither demand significant changes in existing systems nor create a situation in which every given instantiation of a technology would have to incorporate all possible protocols.
The bulk of discussion about ethical concerns, however, involved resource use. It was believed that all connections should be treated in the same way. A problem in one among the "network of networks" comprising the system shouldn't be allowed to bring other networks down. Industrial research organizations were expected to pay their own way, to ensure that commercial concerns didn't degrade the quality of networking for those government agencies and research institutions for whom the Internet was primarily intended during its early years. Those doing research on the network itself were admonished to make sure that their experimentation did not itself cause damage to the network. Access issues, such as pricing policies, were also discussed in ethical terms.
Justifications for ethical standards. The philosophical question of the differences between concepts such as "MUST," "SHALL," and "CAN" is foundational. Use of such terms in programming interacted with discussions about just how compliance with technical standards might be achieved and with the growing formalization of the legal status of decision-making for the Internet. Ultimately the concept of "normative" came to mean that compliance with a given protocol was required, and RFC documents presenting official standards included a section on their normative contexts.
As early as 1989, the argument that the network is a type of infrastructure as important to society and the economy as the road system was used as a justification for insisting on ethical standards in its design and use. The fact that US government funds were essential to its development also meant that adherence to NSF rules regarding unethical and unacceptable activity could be required. Many believed the network should be treated as a common carrier, subject to the regulatory mandates to treat all users in the same manner and not to distort the content in any way. The internationalization of the network, which began in the early 1970s, brought intercultural dimensions of ethical questions into play.
Most of these arguments for insisting upon ethics in Internet design and use remain valid today, and additional political, social, cultural, and economic arguments came into play once the network was commercialized. One justification used in early years is no longer valid, however. In 1991 it was possible to take the position that since no one had to be on the Internet, it was reasonable to require adherence to ethical standards as a quid pro quo for the privilege of being online, but this argument cannot be made a decade into the 21st century.
Mechanisms for ensuring network fairness. The bulk of the techniques for incorporating ethical standards into the network discussed in the RFCs were, of course, technical. These appeared at multiple layers of the network, including at the point of linkage to the network, at gateways between networks, at packet switches. The single technical problem involving equity that received the most treatment up to the point of commercialization was "fair queuing," the issue of how to make sure that the sequencing of messages through the network took place in an equitable manner. Several different approaches were explored. There was concern about whether or not it would be possible to absolutely ensure equity in network use given differences in types of content, uses, and the distance messages have to travel.
Interactions between technology, law, and society. Social policy and behavioral expectations were also treated as mechanisms for ensuring network fairness. The right to have access to encryption, for example, and pricing policies were considered. Specific guidelines were developed to ensure that research on the network wuold not generate ethical problems in the course of experimentation.
Internet designers were aware that it would be possible for bad actors to "game" a system intended to ensure equity. A distinction was drawn between learning how to use a computer system and learning how to use it properly, with the latter including making sure that one is not offending other users. Ethics and etiquette became linked. Computer crime was understood to involve an intertwining of social, legal, and technical matters.
Notions of "good citizenship" came to inform both design decisions and behavioral expectations. The good network citizen shares resources, even when doing so comes at a cost to him- or herself; respects the content, communication, and data of others; respects the resources of others; and complies with protocols and programming rules.
Contributions
This research offers insight into the ways in which specific ethical issues became visible to technical decision-makers as mandates affecting Internet design. It enriches our understanding of the human side of the design process, itself now a model for those building other large-scale socio-technical systems. It is hoped that uses of these findings will help bring technical decision-makers into a common conversation with those in the social sciences and humanities. Finally, the project furthers the development of research methods for analysis of technological discourses with ethical consequences.