The events of September 11th made explicit the effectiveness of asymmetric operations such as terrorism against our country. We know that the democracy and freedom we all take for granted imply vulnerabilities to asymmetric attacks. America has always responded to such challenges with technology. So can technology help us preserve our freedoms while mitigating these vulnerabilities? Can technical tools that facilitate collection, analysis, and dissemination of information throughout the first-responder, law enforcement, intelligence, and military sectors help ensure homeland security? Numerous research and acquisition programs have been initiated in the last 18 months to evaluate and field test such systems. But by their very nature these programs have run into privacy issues, many of them jurisdictional firewalls that are decades old.

Recent examples of such controversial programs include:

The Foresight and Governance Project recently hosted a forum to examine the balance between civil liberties and national security using DARPA’s project as a case study.

The main goals were to:

  • Inform the debate by presenting an accurate view of the technologies being developed.
  • Define the landscape of policy issues that such technologies raise, and the implications for our civil liberties.
  • Discuss perspectives on how to strike the right balance between the social costs and benefits of these technologies.
  • Develop a better understanding of the dynamics of the public debate surrounding such issues.

The discussants were:

John J. Hamre, President and CEO, Center for Strategic & International Studies
Robert Popp, Deputy Director, DARPA/Information Awareness Office
Paul Rosenzweig, Senior Legal Research Fellow, Heritage Foundation
Jim Dempsey, Director, Center for Democracy and Technology
Stuart Taylor, The National Journal
Heather Mac Donald, Manhattan Institute
Christopher Ford, State Department


  • There is a large difference between reality and public perception concerning DAPRA’s Terrorism Information Awareness (TIA) Program, some of this being driven by unrealistic and uninformed portrayals in the media.
  • There is a tendency to see the DARPA technology development work as existing or inevitable. DARPA funds very high risk, high payoff, research so the technologies being proposed or developed under the TIA program may not be deployed for years, or at all.
  • The research portfolio of TIA contains technologies that do not have implications for information policies such as privacy protection (e.g., better peer-to-peer collaboration technologies to help agencies communicate with each other virtually; or foreign language translation technologies to exploit the wealth of information in foreign speech and text).
  • The ground rules and legal/policy frameworks for privacy protection are outdated. These rules were designed for a paper-based economy or centralized databases and not readily applicable to today’s intelligence challenges and emerging information technologies allowing for a more interconnected society.
  • Existing privacy safeguards (many relying on “fair information practices” and social contract) on the civil side may be ineffective when data is transferred to the government side for intelligence purposes. The proposed TIA technologies need to be built consistent with principles such as:

    - notice
    - limitation (no more information accessed or analyzed than needed)
    - reuse (data analyzed for one use should not be passed on for another)
    - retention
    - access (I can see the data you have on me)
    - redress (I can correct errors in that data).

    Redress and access are key issues that must be solved. However, some of these principles may not be readily transferable to the intelligence/law enforcement regime, in which case adequate substitutes must be found.
  • Unlike subject-based queries, there is a lack of a policy framework or set of legal protocols to deal with the type of pattern-based queries being tested or contemplated under the TIA program. A model analogous to “probable cause” coupled with a privacy protection approach such as “selective revelation” appears to be a reasonable starting point to justify such searches.
  • Both the creators and the users of technologies such as those being developed in TIA must be sensitized to civil liberties issues (training may be necessary).
  • Safeguards should be “hardwired” into these technologies whenever possible, including tamper-proof audit trails, policy and business rule enforcement, encryption, etc.
  • The problem of false positives is a challenge and concern – both technically and from a policy perspective – for any predictive system such as TIA and will need to be contended with.
  • DARPA, and other developers of such systems, need to be open and forthright about the technologies under development and their capabilities in order to inform public debate.
  • When confronted with these types of technologies, the press will have tendencies to imagine the worse, sanctify the status quo, leave out balancing opinions, and not explore alternative solutions (as well as their costs and benefits).
  • Assume technological and scientific illiteracy on the part of the media. Even the best and most well-intentioned reporters will have problems understanding these technologies. DARPA cannot change the press; DARPA must change its strategies for dealing with the press. In attempting to inform the press or public, use examples of what the technologies do, if possible.
  • Privacy is a misunderstood concept and makes discussions more difficult. It may be better to focus on anonymity and people’s “expectations” about anonymity.

The dialogue was moderated by Dave Rejeski, Director of the Foresight and Governance Project at the Wilson Center. The event was organized by Victoria Stavridou, SRI International, and Jim Kadtke, Office of Senator John Warner.