September 2, 2003

The Honorable Ted Stevens
Chairman, Senate Appropriations Committee
SH-522 Hart Senate Office Building
Washington, DC 20510-0201

Dear Chairman Stevens:

As an association comprised of 200 university computer science departments, 25 industrial research labs and six affiliated societies representing nearly 150,000 computing professionals, we ask that you and your colleagues on the conference committee for the FY 2004 Defense Appropriations Act (S. 1382, H.R. 2658) support language found in the House version of the bill (sec. 8124) that would prohibit the deployment of any 'Terrorism Information Awareness' system without Congressional approval, but which would allow critical national security research to move forward. We have great concerns about alternate language in the Senate version of the bill (sec. 8120) that would not only prohibit deployment of any TIA system, but would also prohibit fundamental research in datamining, privacy and security already underway as part of TIA-related programs at DARPA.

The TIA research program has been the focus of much criticism and concern from within the computing research community. The program, an attempt to 'design a prototype network that integrates innovative information technologies for detecting and preempting foreign terrorist activities against Americans,' has come under fire by many who see the eventual development of such a network as a serious threat to American civil liberties and security. A TIA system could potentially be ripe for misuse by authorized users, could be prone to false positives (in which innocent individuals might be identified as candidate terrorists), could itself be a target of attack or make the systems with which it interacts less secure and, ultimately, might not help the nation protect itself from terrorism.

As an association of computing researchers, CRA has shared many of these concerns through the DARPA advisory process and with policymakers in Congress. However, we wish to emphasize that the technologies proposed for research under TIA could have valid uses in many other contexts, including predicting failure of safety-critical components, identifying fraud in contracting, and identifying suspicious transfers of controlled materials. Many other uses may be discovered as the technology matures, including in counter-terrorism. The military and the country have repeatedly benefited from the technological advantages that result from research into difficult computing problems -- including information fusion, improved privacy technologies, and machine-learning algorithms for data mining. These are technologies that have been identified by many, including the National Research Council as worthy of further study. The problems that need solving are, indeed, sufficiently hard ('DARPA-hard') as to be worth the time and investment as independent research thrusts.

Therefore, CRA supports continued research in this area and would oppose a moratorium as proposed in S. 1382. For the same reason, we oppose the significant cut to 'Asymmetric Threat Programs' under the 'Computing Systems and Communications Technology' line included in S. 1382. We ask that you and your colleagues on the conference committee adopt sec. 8124 of H.R. 2658 and restore funding for these important research areas when you assemble the final conference report.

Finally, we commend you and your colleagues on the committee for your continued support of DOD science and technology and for your willingness to look to new technologies to aid in securing the nation's defense. As these technologies mature, the need for strong congressional oversight over their development and deployment will increase significantly. We stand ready to assist you and your staff in that oversight process. If we can be of any assistance as the conference moves forward, please do not hesitate to ask.

Sincerely,

James D. Foley
Chair
Computing Research Association

Related Links

Submitted by pch on 2003-09-02