Peeping Hals

The field of artificial intelligence, broadly defined as the study and practice of designing intelligent agents, is at least six decades into its existence as a formal discipline. Sometimes called “computational” or “synthetic” intelligence, AI borrows from and informs a wide variety of subjects, including philosophy, psychology, linguistics, neuroscience, statistics, economics, and law. Techniques of AI underpin all manner of industrial and consumer applicationsfrom the complex neural nets used in data mining, down to the ‘fuzzy logic’ used by commercial washers and driers. Insofar as “the issues of AI are directly related to [the] self-image of human beings”, 4 and because the central projects and techniques of AI can often be articulated in lay language, few shy away from offering their insights and critiques. This essay explores and updates a particular criticismthe long-standing claim that certain techniques and applications of AI violate human privacyand discusses whether (U.S.) privacy law is adequately positioned to respond. Historically, AI can be said to threaten privacy according to a specific pattern: AI substitutes for humans at various stages of observation or surveillance, allowing such activity to reach a previously impracticable scale. Whereas once telephonic surveillance required one listener per phone call, the development of voice recognition technology permits the substitution of a computer capable of monitoring thousands of calls simultaneously. Whereas once hundreds of intelligence analysts might be required to pour over field records in search of connections, AI knowledge management techniques automatically spot patterns