The Digital Case Officer: Reimagining Espionage with AI (scsp222.substack.com)

The Special Competitive Studies Project released a report and podcast episode arguing that generative, multimodal and agentic AI — combined with cheaper, domain-specific models and large-scale data analysis — are poised to upend human intelligence (HUMINT). SCSP coins the “Digital Case Officer”: AI systems that can ingest vast, multimodal datasets and autonomously target, profile, groom, recruit and advise human sources at scale. The report warns that Ubiquitous Technical Surveillance (phones, cameras, online traces) is already eroding traditional tradecraft and that adversaries (notably the PRC) are moving fast to deploy these tools, making legacy methods riskier while leaving some tasks (e.g., penetrating air-gapped systems, discerning intent) still dependent on human officers. Technically, the proposal emphasizes agentic systems for automated targeting, psychological profiling from digital footprints, and simultaneous management of hundreds of development threads — all under a human-machine team model where AI handles scale and data synthesis and people retain final judgment. It outlines governance and operational safeguards: formal CIA/ODNI guidance, Meaningful Human Control and human-in-the-loop checkpoints for high-risk actions, continuous red-teaming, an interagency “HUMINT-as-a-Service” model to avoid fragmented toolsets, revised tradecraft training, and named accountability officers with reporting requirements. The piece frames swift, policy-attuned adoption as essential to retain strategic advantage while mitigating legal, ethical and operational risks.