Restoring Transparency to Automated Authority

From AcaWiki
Jump to: navigation, search

Citation: Frank Pasquale Restoring Transparency to Automated Authority.
Internet Archive Scholar (search for fulltext): Restoring Transparency to Automated Authority
Download: http://ssrn.com/abstract=1762766
Tagged:

Summary

"Concern over gaming provoked a shift away from transparency as a legitimation strategy; instead, ironclad secrecy has been pursued."

"The move from legitimation-via-transparency to reassurance-via-secrecy has profoundly troubling implications for the foundations of social order in the information age."

"Unfortunately, the law is presently stacking the deck against accountability for automated authorities." Ex: data copyright, anticircumvention, trade secret, EULAs.

Examples from health, web search, and finance.

"Just as the ―fair use‖ doctrine has deterred the overpropertization of expression, generally recognized fair information practices should include large and powerful data holders‘ obligation to surrender some sample of their data to entities entrusted to audit and assess the data holders‘ activities."

Conclusion:

Laws promoting transparency have shed some light on troubling practices.However, new automated authorities are often so complex that merely revealing them will not solve the problems discussed above. Transparency should be a first step toward an intelligible society, where leading firms‘ critical decisions can be understood not merely by their own engineers and mathematicians, but also by risk managers and regulators.However well an ―invisible hand coordinates economic activity generally, markets depend on reliable information about the practices of core firms that finance, rank, and rate entities in the rest of the economy. Brandishing quasi-governmental authority to determine which enterprises are funded and found, they need to be held to a higher standard than the average firm.

Theoretical and Practical Relevance

Mentions open source once:

Longman asserts that the ―largest and most successful example of digital medicine is an open-source program called VistA, and contrasts it with proprietary systems ―written by software developers who are far removed from the realities of practicing medicine. Longman worries that several proprietary systems increase the chance of medical error due to restrictive licensing agreements which prohibit users from revealing system problems.

Implies revelation through patenting previously was adequate for transparency of computational intermediaries. Seems highly unlikely.