Sunday, October 29, 2006

Policy implications of the Cognitive Radio metaphor

Practitioners who use the cognitive radio metaphor understand that radios are not conscious in the ways that humans are; they use this term to inspire research, the results of which then stand on its own merits, separately from the guiding metaphor.

However, policy and business decision makers may not have such an informed and subtle perspective. “Cognitive radio” is such a powerful image that they may be tempted to take the metaphor at face value, and make inferences that reach beyond the bounds of the model.

For example, they may infer that cognitive radio systems will be as flexible and sophisticated as the cognitive system they know best: human beings. This inflates expectations. The “AI winter” of the late 1970s was due, in part, to the dearth of promised results despite substantial funding since the mid-1950s. While one cannot directly compare a new endeavor like cognitive radio with the relatively mature AI field in the Seventies, the lesson of expectation management is an obvious one.

The terminology may also lead decision makers to over-estimate the risks of cognitive radio. The Radios-as-People model not only entails that the devices will be smart, flexible, and intelligent; it also intimates that some will be malicious, devious, malevolent, deceitful, treacherous, untrustworthy, dangerous or destructive. Cognitive radio technology may be unfairly judged to be a severe security risk simply on the basis of the connotations its name evokes.

The “radio” part of the conceptual blend invites regulation where it might otherwise not be contemplated. Regulators have taken a largely hands-off approach to computing. However, governments have been regulating radio for almost a century; they are not only comfortable doing so, many feel it is their duty. While the blend of “cognitive” and “radio” may bring computing’s unregulated regime to wireless, it may conversely bring radio regulation to computing.

“Cognitive” implies informed behavior; cognitive radio thus raises the question of regulating radio behavior. Radio regulation to date has specified parameters so simple that the term “behavior” is scarcely merited. This regulation has applied almost exclusively to transmitters; receiver standards have been rare, and those that exist specify passive parameters like selectivity and spurious rejection [1]. Regulating cognitive radio thus raises challenges for agencies with scant experience with receiver standards, let alone software behavior verification.

---------

[1] National Telecommunications and Information Administration of the U.S. Department Of Commerce (2003) , Receiver Spectrum Standards, Phase 1 – Summary of Research into Existing Standards, NTIA Report 03-404, at http://www.ntia.doc.gov/osmhome/reports/ntia03-404/NTIAREPORT03-404.pdf

1 comment:

suze woolf said...

It's because the perfectly good Anglo-Saxon "smart" has become debased by garage door openers and cell phones. Smart doesn't sound nearly so worthy of reulation.