December 7, 2010; Source: | When the nonprofit Constitution Project issues a report that affects how nonprofits and government interact, it’s probably worth looking at.

Its newest report, Principles for Government Data Mining: Preserving Civil Liberties in the Information Age [PDF], the Constitution Project worries that government data mining “can offer significant benefits, but without adequate processes and controls, it can encroach on constitutional rights and values—including privacy, freedom of expression, due process, and equal protection.”

Data mining refers the automated process of finding correlations or patterns among dozens of fields in large relational databases.

Besides citing the obvious examples of the government’s controversial anti-terrorism lists, the Project presages the WikiLeaks controversy and warns that “Rogue government employees can abuse database access and look for information on the famous or infamous—as occurred with the 2008 presidential candidates.” With hindsight, they might have added military reports and classified State Department cables as well.

But some types of data mining, “such as those that improve program efficiency and evaluate performance are very worthwhile.” Nonprofits are probably slowly becoming aware that data mining is being used to detect tax fraud and to spot and investigate potential misuse of economic stimulus funds.

For example, the Recovery Accountability and Transparency Board, the so-called RAT Board, has been using anti-terrorism-like data mining techniques to identify not only improper disbursements such as overpayments, but also “hard-to-find conflicts of interest between recipients and the contracts they have bid on.”

The Constitution Project implies that the definition of data mining used by government isn’t broad or inclusive enough, allowing some data mining initiatives to escape the reporting and disclosure required by the 2007 Data Mining Reporting Act. Nonetheless, despite government’s own reporting and classification deficiencies, nonprofits should assume that their contract information is subject to the pattern analysis that underlies much of data mining technology, and if they mess up, they might find their errors dug out and brought to the surface by federal data mining software—whether they like it or not.—Rick Cohen