Aldrich Ames: The Insider Who Sold the Network

Aldrich Ames was a CIA officer in the Soviet and East European Division, working in counterintelligence, which meant he had access to the identities of the people the United States was running inside the Soviet system.

In April 1985 he walked into the Soviet embassy in Washington and offered to sell.

He initially intended a single transaction — fifty thousand dollars, in exchange for the names of two double agents he believed were Soviet-controlled anyway. Then he provided the rest. Over the following months he handed over essentially the entire roster.

At least ten of those people were executed. More were imprisoned. The collapse was so comprehensive and so sudden that the CIA knew within a year that it had a penetration, and spent nine years failing to find him.

The insider-threat literature — Shaw and Fischer's studies among others — describes a standard sequence, and Ames fits it.

Personal financial pressure: an expensive divorce, a new Colombian wife with expensive tastes, debts.

Perceived professional grievance: Ames considered himself undervalued, had received indifferent evaluations, had a drinking problem that had been noticed and not acted on, and believed the agency's work was largely futile.

And moral disengagement, in Bandura's sense: the agency was incompetent and deserved it; the assets were doomed anyway; the whole enterprise was a game; the money was owed to him. Every one of these is a recognized mechanism for maintaining a self-concept while doing something incompatible with it.

The detection failure is the part worth studying, and it is not primarily technical.

Ames's spending was flagrant. He paid five hundred and forty thousand dollars in cash for a house in Arlington. He drove a Jaguar. He wore suits nobody on a government salary could afford. Colleagues noticed and made jokes about it; the explanation offered — his wife's family money in Colombia — was accepted for years without being checked.

Two failures compound here. Base-rate neglect: penetrations are rare, so the prior is low, and a low prior means anomalies get explained away. And the reluctance to suspect a colleague, which is a social cost that falls entirely on the person raising the concern.

He passed polygraph examinations in 1986 and 1991, which is the subject of chapter 115.

What eventually caught him was a joint CIA-FBI team doing exactly the mundane thing: correlating his financial records against his contacts and his travel, and finding that deposits followed meetings.