Old-School Spycraft Could Make a Comeback as AI Undermines Trust

Why It Matters

The shift forces intelligence agencies to balance rapid AI adoption with renewed reliance on secure, low‑tech human methods, reshaping national‑security operations and procurement priorities.

Key Takeaways

• AI deepfakes erode trust in electronic communications.
• Dead drops and brush passes regain operational relevance.
• HUMINT remains essential despite AI-driven data collection.
• CIA modernizes tech procurement while emphasizing human sources.
• AI can aid case officers in persuasive messaging.

Pulse Analysis

The rapid diffusion of generative AI has turned digital channels into a battlefield of authenticity. Deep‑fake videos, synthetic voice messages and AI‑crafted text can be produced at scale, flooding intelligence analysts with indistinguishable noise. As Thomas Mulligan notes in his RAND‑backed study, this erosion of trust forces agencies to question every electronic transmission, because a single fabricated packet can mislead a case officer or trigger a costly operation. The resulting credibility gap is not merely technical; it reshapes risk assessments and forces a reevaluation of how agencies validate source information.

Against that backdrop, the oldest espionage tools—dead drops, brush passes, and face‑to‑face briefings—are experiencing a quiet renaissance. A dead drop, hidden in a mundane location, lets an officer confirm a source’s physical presence without exposing either party to digital interception. Brush passes, executed in crowded public spaces, provide rapid, low‑tech exchanges that leave no electronic trail. These methods reintroduce a human verification layer that AI cannot replicate, allowing operatives to separate genuine pleas for help from algorithm‑generated scams and to preserve operational security in hostile environments.

The CIA’s recent overhaul of its acquisition framework reflects a dual strategy: accelerate AI integration while safeguarding the human element that underpins HUMINT. By streamlining contracts for classified‑grade AI tools, the agency hopes to boost analytical speed, yet Mulligan stresses that persuasion, relationship building, and on‑the‑ground judgment remain irreplaceable. This balanced approach signals to contractors and policymakers that AI will augment, not supplant, the case officer’s craft. For the intelligence community, the lesson is clear—embrace technological advantage, but retain the analog safeguards that have defined espionage for centuries.