On 31 January the National Science Foundation (NSF) and the OECD held a joint conference on “social and economic factors shaping the future of the Internet.” Attendance was restricted to about 20 full participants selected by OECD/NSF and another 30 attendees who were allowed to ask questions.
For NSF, the meeting promoted their effort to incorporate “social and economic factors” into the research around their Global Environment for Network Innovation (GENI) initiative. GENI is a major new NSF initiative to fund a “clean slate” redesign of the Internet. The intellectual driver of this initiative seems to be David Clark, who was also one of the leading protocol architects of the old Internet. Suzi Iacono, the NSF program officer who concentrates on the social, economic and behavioral aspects of information systems within the NSF's CISE division, believes that GENI can create a “testbed” that will allow social scientists to experiment with the way various protocol or network designs interact with social factors.
My first observation about this meeting is that aside from David Clark's always-interesting ruminations on what problems a clean-slate resdesign of the Internet might involve, very few new ideas were bruited. Almost all of the discussion revolved around the social, economic and political problems of the “old” Internet. More importantly, I wonder whether the desire to link analysis and understanding of social problems to the engineering or redesign of a new Internet is unambiguously a good thing.
The tension between the desire to “design” a new Internet and at the same time take into account social and economic factors was confronted directly at one point. Clark was asked by Harvard's Jonathan Zittrain about the ITU's Next Generation Network (NGN) initiative. An excellent question — isn't that also an attempt to do a clean slate redesign? Clark answered that the ITU's effort was “doomed” because “people are watching.” By that he meant that the basic design issues that would really have an effect on social factors would be “torn apart by advocacy.” In other words, assuming that social effects can be controlled by means of technical design (a very big and mostly false assumption) interested social actors will not stand passively by and let those design decisions be made without their input.
Now if the outcomes of NGN will be torn apart by advocacy, why not also the results of GENI? All Clark could say was that one should “design the levers of control so that you've got the right balance of power over the controls.” But won't those design decisions be influenced by existing inequalities in political and economic power?
My point is not that we should throw up our hands and say that nothing can be done or that none of this should be carefully studied. Rather, my target is the assumption that one can “design” large-scale technical systems in ways that avoid future political and economic problems. If engineers try to take “social and economic factors” fully into account as they design technical systems, it is inevitable that the design will reflect a status quo oriented equilibrium among existing vested interests. The technological design will embed rules and restrictions based on protection of today's business models and political interests.
Engineers and scientists can be much more revolutionary (in both a good and bad sense) by solving technical problems in complete disregard of their social and economic consequences. Or by paying attention to some social factors and disregarding others. This is what as happened with the original Internet. It is the job of social, political and economic institutions, not technologists, to respond to the problems of technological innovation as they arise. It is a false and technocratic assumption that engineers and social scientists can embed these decisions in technological design. Indeed, this widespread belief that beneficial social consequences can somehow be “embedded” in technological design is one of the truly bad consequences of the Lessig meme that “code is law.”