Data Collection: Lessons of Cost-benefit Analysis, Skepticism, and Legal Transparency

Data Collection: Lessons of Cost-Benef‌it Analysis,

Skepticism, and Legal Transparency

James X. Dempsey*

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

I. ENDING COLLECTION PROGRAMS BASED ON A COST-BENEFIT ANALYSIS . 127

II. WHEN PRIVACY IS AT STAKE, EFFICACY SHOULD ALWAYS BE AN ISSUE. 129

III. A HEALTHY SKEPTICISM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

IV. MAKING EFFICACY ANALYSIS ROUTINE . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

V. TRANSPARENCY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

CONCLUSION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

INTRODUCTION

In the realm of electronic surveillance, at least three lessons can be drawn from

the experience of the last twenty years: (1) policymakers and the intelligence

community can decide to not collect, or to stop collecting, information when

its value does not outweigh the intrusiveness of its collection, even leaving

aside questions of constitutionality or statutory authority; (2) policymakers,

intelligence agency off‌icials, and members of the public should be skeptical of

claims—coming from both the intelligence community and the tech industry—

about the value of big data analytics; and (3) intelligence collection programs can

be described with some detail in legislation without compromising their effec-

tiveness. For these lessons, I draw upon the experience of the now-defunct teleph-

ony metadata program under Section 215 of the PATRIOT Act and the much

more successful collection activities under Section 702 of the Foreign

Intelligence Surveillance Act.

I. ENDING COLLECTION PROGRAMS BASED ON A COST-BENEFIT ANALYSIS

One of the hard decisions in the intelligence f‌ield (a f‌ield characterized by hard

decisions) is the choice to not collect certain information. At one level, intelli-

gence is all about selection. It requires the constant funneling of oceans of infor-

mation through f‌iner and f‌iner sieves to reach the point of knowledge or insight

offered with some conf‌idence. In the digital age, however, there is so much infor-

mation available with such relative ease, with such powerful tools available to

store and analyze it and such a faith in technology, that there may be a tendency

to collect it “just in case.” Once a collection program begins, it can be very

* Lecturer, UC Berkeley School of Law; Senior Policy Advisor, Program on Geopolitics, Technology and

Governance, Stanford Cyber Policy Center; Member, Privacy and Civil Liberties Oversight Board (2012-

2017).

127