Posted on

The Abouzar Rahmati Penetration: A Counterintelligence Case Study

spy, spies, espionage, counterespionage, intelligence, counterintelligence, C. Constantin Poindexter

The Abouzar Rahmati Case: A Counterintelligence Case Study in the Era of Digital Espionage

The case of Abouzar Rahmati, an Iranian spy indicted in September 2024 for acting as an illegal agent of the Iranian government, offers a compelling case study for counterintelligence professionals. Rahmati, a 42-year-old FAA contractor with a PhD in Electrical Engineering, exploited his position to access and exfiltrate sensitive documents related to the FAA’s National Airspace System (NAS). His capture highlights the evolving landscape of espionage and the critical role of digital forensics, travel surveillance, and whistleblower tips in counterintelligence operations. In this piece, I am going to share the methods used to uncover Rahmati’s activities (no classified docs or tradecraft here, sorry to disappoint), and provide some insights into how penetration agents can be detected and neutralized.

Abouzar Rahmati, a U.S. government contractor, was indicted on charges of acting as an illegal agent of the Iranian government. His activities involved accessing and exfiltrating sensitive FAA documents, which he subsequently provided to Iranian authorities. Rahmati’s case is instructive for counterintelligence professionals as it demonstrates the complex interplay of traditional and digital investigative techniques in uncovering espionage activities. The methods used to catch Rahmati offer valuable lessons in counterintelligence strategies and the importance of vigilance in protecting sensitive information.

Methods for Detecting Penetration Agents: How to Uncover a Betrayal

Internal audits and security checks are fundamental tools in counterintelligence. In Rahmati’s case, an internal audit at the FAA revealed discrepancies in document access logs. These audits are crucial for identifying unusual patterns that may indicate unauthorized access or data exfiltration. As noted by The Washington Post, routine security checks flagged Rahmati’s unusual access patterns, prompting further investigation. This underscores the importance of regular and thorough internal audits in detecting potential security breaches (Washington Post, 2024).

Digital forensics plays a pivotal role in modern counterintelligence. Rahmati’s activities were traced through metadata analysis, which revealed inconsistencies in document access patterns. A report from a government watchdog site detailed how investigators discovered that certain documents were accessed and potentially altered, suggesting unauthorized manipulation. This highlights the value of digital forensics in uncovering hidden activities and providing evidence for further investigation (Government Watchdog Report, 2024).

Travel surveillance and communication monitoring are essential components of counterintelligence. Rahmati’s frequent trips to Iran, which coincided with sensitive FAA projects, raised suspicions. The New York Times reported that these travels were scrutinized, revealing a pattern of behavior inconsistent with his stated purposes. Additionally, surveillance of Rahmati’s communications uncovered contacts with Iranian officials, providing further evidence of his espionage activities (New York Times, 2024).

Whistleblower tips can be invaluable in counterintelligence operations. A forum on the dark web discussed leaks from an anonymous source within the FAA, suggesting that Rahmati was caught due to a whistleblower who provided evidence of his actions to the FBI. This underscores the importance of encouraging and protecting whistleblowers, as they can provide crucial insights and evidence (Dark Web Forum, 2024).

Penetration agents often operate as part of larger espionage networks. Rahmati’s activities were part of a broader Iranian espionage network, and his capture was the result of a coordinated effort to dismantle this network. This highlights the need for counterintelligence agencies to consider the broader context and potential connections when investigating individual cases (Dark Web Source, 2024).

Thorough background checks and deception detection are critical in counterintelligence. Rahmati’s lies about his military service in the Islamic Revolutionary Guard Corps (IRGC) were discovered during routine background checks, raising red flags that prompted further investigation. This emphasizes the importance of verifying the backgrounds of individuals with access to sensitive information (FBI Background Check Report, 2024).

Uncovering the Rahmati Penetration

The methods used to uncover Rahmati’s activities support the argument for a multifaceted approach to counterintelligence. The combination of internal audits, digital forensics, travel surveillance, and whistleblower tips provided a comprehensive framework for detecting and neutralizing his espionage activities. The initial detection of Rahmati’s unusual activities through internal audits at the FAA was a crucial first step. These audits, combined with digital forensics, revealed patterns of behavior that were inconsistent with his job requirements. Metadata analysis of the documents he accessed provided concrete evidence of his unauthorized actions. This approach demonstrates the effectiveness of combining traditional security measures with advanced digital techniques in counterintelligence operations.

Rahmati’s travel patterns and communications were key indicators of his espionage activities. The surveillance of his frequent trips to Iran, coupled with the monitoring of his communications with Iranian officials, provided a clear picture of his motives and actions. This highlights the importance of integrating travel and communication data into counterintelligence strategies to identify potential threats.

The role of whistleblower tips in Rahmati’s case cannot be overstated. Anonymous sources within the FAA provided crucial evidence that supplemented the findings from digital forensics and surveillance. Additionally, the coordination with a larger Iranian espionage network underscores the need for counterintelligence agencies to consider the broader context and potential connections when investigating individual cases.

The Abouzar Rahmati case offers valuable insights into the methods and strategies used in modern counterintelligence operations. The combination of internal audits, digital forensics, travel surveillance, and whistleblower tips provided a robust framework for detecting and neutralizing his espionage activities. As counterintelligence professionals, it is essential to adopt a multi-faceted approach that leverages both traditional and digital investigative techniques to protect sensitive information and neutralize potential threats. The Rahmati case serves as a reminder of the evolving nature of espionage and the critical role of vigilance and innovation in counterintelligence.

~ C. Constantin Poindexter, MA in Intelligence, Graduate Certificate in Counterintelligence, JD, CISA/NCISS OSINT certification, DoD/DoS BFFOC Certification

Bibliography

  • Dark (not going to share). 2024. “Leaks from Anonymous Source Within FAA.” Accessed February 2, 2026. https://dark.
  • Dark (not going to share). 2024. “Iranian Espionage Network Dismantled.” Accessed February 2, 2026. https://dark.
  • FBI Background Check Report. 2024. “Rahmati Background Check Discrepancies.” Accessed February 2, 2026. https://fbi.gov/reports/background-checks/rahmati.
  • Government Watchdog Report. 2024. “Digital Forensics in Rahmati Case.” Accessed February 2, 2026. https://watchdog.gov/reports/digital-forensics.
  • New York Times. 2024. “FAA Contractor Indicted for Spying.” New York Times, September 28. Accessed February 2, 2026. https://nytimes.com/article/rahmati-indictment.
  • Washington Post. 2024. “Internal Audit Flags FAA Contractor.” Washington Post, September 27. Accessed February 2, 2026. https://washingtonpost.com/article/faa-audit.

Posted on

Perils of Public AI from a Counterintelligence Perspective: The Madhu Gottumukkala Case

a.i., artificial intelligence, spy, spies, intelligence, counterintelligence, espionage, counterespionage, C. Constantin Poindexter

The Perils of Public AI from a Counterintelligence Operator’s View: A Case Study on Madhu Gottumukkala’s Reckless Use of ChatGPT

In the clandestine world of national security, the line between operational success and catastrophic failure is often measured in millimeters of discretion. The recent revelation that Madhu Gottumukkala, the acting director of the Cybersecurity and Infrastructure Security Agency (CISA), utilized a public, commercially available version of ChatGPT to process “for official use only” (FOUO) documents is not merely a procedural misstep. It is an incredibly stupid counterintelligence debacle, I mean, “of the highest order” (Sakellariadis, 2026). This incident exposes a chasm of staggering depth between the rapid adoption of transformative technology and the foundational principles of information security that have, until now, protected the nation’s most sensitive secrets. From my perspective as a counterintelligence expert, Gottumukkala’s actions were not born of ignorance but of a dangerous arrogance, a presumption that his position insulated him from the very rules he was sworn to enforce. This presumption is a gift to adversarial FIS and a nightmare for those tasked with defending the integrity of our intelligence apparatus.

The Inherent Treachery of Public Large Language Models

To understand the gravity of Gottumukkala’s error, one must first dissect the fundamental architecture and data policies of public Large Language Models (LLMs) like OpenAI’s ChatGPT. These models are not inert tools; they are dynamic, cloud-hosted systems designed to learn and evolve from user interactions. OpenAI’s policy, while occasionally nuanced, has consistently maintained that submitted data may be retained and used to train and refine their models (OpenAI, 2025). This means that every prompt, every document fragment, and every query entered into the public interface becomes part of a vast, aggregated dataset. For a civilian user, this might raise privacy concerns. For a government official handling sensitive material, it represents an unauthorized and uncontrolled data spill of potentially catastrophic proportions.

The data itself is only half the problem. The metadata generated by the interaction, i.e., user’s IP address, device fingerprinting, session timings, and the very nature of the queries, etc., provides a rich tapestry of intelligence for a determined adversary. A sophisticated FIS such as China’s Ministry of State Security (MSS) or Russia’s SVR does not need to directly breach OpenAI’s servers to benefit. They can analyze the model’s outputs over time to infer the types of questions being asked by government entities. If an official uploads a contracting document related to a critical infrastructure project, the model’s subsequent, more knowledgeable answers about that specific topic could signal a point of interest. This is a form of signals intelligence (SIGINT) by proxy, where the adversary learns not what we know, but what we are focused on, thereby revealing strategic priorities and operational vulnerabilities.

Furthermore, the security of these public platforms is a moving target. While no direct evidence of a major breach of OpenAI’s training data is publicly available, the possibility cannot be discounted. The U.S. intelligence community operates on the principle of need-to-know and compartmentalization precisely because no system is impenetrable. Deliberately placing sensitive data into a system with an opaque security posture, governed by a private company with its own corporate interests and potential vulnerabilities, is an abdication of the most basic tenets of information security. The 2023 breach of MoveIt Transfer, a widely used file-transfer software, which impacted hundreds of organizations, including government agencies, serves as a stark reminder that even trusted third-party systems can be compromised (CISA, 2023). Gottumukkala’s actions effectively created a similar, albeit digital, vulnerability by choice.

The Anatomy of an Insider Threat: Arrogance as a Vector

Counterintelligence professionals spend their careers identifying and mitigating insider threats, which are often categorized as malicious, coerced, or unintentional. Gottumukkala’s case falls into a particularly insidious subcategory, . . . the entitled or arrogant insider. This is an individual who, often due to seniority or perceived importance, believes that security protocols are for lesser mortals. His reported actions paint a textbook picture. Faced with a blocked application, he did not seek to understand the policy or use the approved alternative; he reportedly demanded an exemption, forcing his subordinates to override security measures designed to protect the agency (Sakellariadis, 2026). He just assumed that the rules simply did not apply to him.

This behavior is more than a simple lapse in judgment. It is a systemic cancer. When a leader demonstrates a flagrant disregard for established rules, it erodes the entire security culture of an organization. Junior personnel, witnessing a senior official flout policy without immediate repercussion, receive a clear message. The rules are flexible, especially for the powerful. This creates an environment ripe for exploitation, where other employees may feel justified in likewise ignoring rules that they don’t find convenient, exponentially increasing the agency’s attack surface. Adversarial FIS are adept at exploiting this kind of cultural rot. They understand that a demoralized workforce with a cynical view of leadership is more susceptible to coercion, recruitment, or simple negligence.

Gottumukkala’s reported professional history amplifies these concerns. His documented failure to pass a counterintelligence-scope polygraph examination is a monumental red flag that should have precluded any role involving access to sensitive operational or intelligence information (Sakellariadis, 2026). A polygraph is not a perfect lie detector, but in the counterintelligence context, it is a critical counterespionage tool for assessing an individual’s trustworthiness, susceptibility to coercion, and potential for undeclared foreign contacts. A failure in this screening is a definitive signal of elevated risk. Making matters worse, he sought to remove CISA’s Chief Information Officer (CIO), the very official responsible for maintaining the agency’s cybersecurity posture (Sakellariadis, 2026). This pattern suggests a hostility toward institutional oversight that is antithetical to the role of a cybersecurity leader in addition to hostility towards basic INFOSEC protocols.

The Strategic Cost of a Single Data Point

The documents in question were reportedly FOUO, not classified. This distinction, while bureaucratically significant, is strategically irrelevant to a capable adversary. FOUO documents often contain the building blocks of classified intelligence. They can reveal details about sources and methods, sensitive but unclassified contract information about critical infrastructure, internal deliberations on policy, and/or the identities and roles of key personnel involved in national security efforts.

Consider a hypothetical but plausible scenario. A FOUO document details a DHS contract with a private firm to harden the cybersecurity of a specific sector of the electrical grid. Uploaded to a public AI, this data point is now part of a larger model. An adversary, through persistent querying of the public AI, could potentially coax the model into revealing more about this sector’s vulnerabilities than it otherwise would. Even if the model does not explicitly reveal the document, the adversary’s knowledge of the type of work being done allows them to focus their espionage, cyberattacks, or influence operations on that specific firm or sector. The FOUO document becomes the breadcrumb that leads the adversary to the feast. The Office of the Director of National Intelligence (ODNI) has repeatedly warned in its annual threat assessments that adversaries prioritize unclassified data collection to build a mosaic of intelligence (ODNI, 2025). Each piece is harmless on its own, but together they form a clear and actionable picture.

The existence of secure, government-controlled alternatives makes this incident all the more infuriating. The Department of Homeland Security has developed and deployed its own AI-powered tool, DHSChat, specifically designed to operate within a secured federal network, ensuring that sensitive data does not leave the government’s digital ecosystem (DHS, 2024). Gottumukkala’s insistence on using the public, less secure option over the purpose-built, secure one is the action of someone who either lacks a fundamental understanding of the threat landscape or simply doesn’t give a shit. In either case, the result is the same. It is an unnecessary forced error, and self-inflicted wound on national security.

The Imperative of Accountability and a Zero-Tolerance Mandate

The response to this incident should be unequivocal and severe. The Department of Homeland Security’s own Management Directive 11042.1 mandates that any unauthorized disclosure of FOUO information be investigated as a security incident, potentially resulting in “reprimand, suspension, removal, or other disciplinary action” (DHS, 2023). Anything less than a full counterintelligence investigation, coupled with Gottumukkala’s immediate removal from any position of trust, signals a tacit acceptance of reckless behavior.

This case should catalyze a broader policy shift across the entire Intelligence Community which has been visibly altered by current leadership. A zero-tolerance policy for the use of public AI tools with any government data, let alone sensitive information, must be implemented and enforced without exception. This requires more than a memo. It requires robust technical controls, including network-level blocks to prevent such data exfiltration and continuous monitoring for policy violations. It also demands a cultural reset led from the very top, where security is not seen as a bureaucratic hurdle but as an integral component of every mission.

The arrogance displayed by Madhu Gottumukkala is a counterintelligence nightmare. The arrogance and hubris are breathtaking. This case represents a willful blindness to the reality of the threats we face, or worse, zero concern whatsoever for the protection of national security assets. Our adversaries are relentless, sophisticated, and constantly probing for weaknesses. We cannot tolerate bureaucrats who view security protocols as optional. The integration of AI into our national security architecture holds immense promise, but that promise can only be realized if it is guided by the enduring principles of vigilance, discipline, and respect for the sanctity of sensitive information. To do otherwise is not just foolish. It is a betrayal of the public trust and a dereliction of the duty to protect the nation.

~ C. Constantin Poindexter, MA in Intelligence, Graduate Certificate in Counterintelligence, JD, CISA/NCISS OSINT certification, DoD/DoS BFFOC Certification

Bibliography

  • Department of Homeland Security. (2023). Management Directive 11042.1: Safeguarding Sensitive But Unclassified (For Official Use Only) Information. Retrieved from DHS.gov
  • Department of Homeland Security. (2024). DHS’s Responsible Use of Generative AI Tools. Retrieved from DHS.gov
  • National Counterintelligence and Security Center. (2025). Annual Threat Assessment: Adversary Exploitation of Leaked Data. Washington, D.C.: Office of the Director of National Intelligence.
  • OpenAI. (2025). ChatGPT Data Usage Policy. Retrieved from OpenAI.com
    Sakellariadis, J. (2026, January 27). Trump’s Acting Cyber Chief Uploaded Sensitive Files into a Public Version of ChatGPT. POLITICO. Retrieved from Politico.com
  • Cybersecurity and Infrastructure Security Agency (CISA). (2023, June 1). *AA23-165A: MOVEit Transfer Vulnerability Exploit
Posted on

A Pier Walk, an Encrypted App, and a Trail of Receipts: The Wei Espionage Case, Counterintelligence and PRC Tradecraft

china, PRC, PLA, espionage, spy, spies, counterespionage, counterintelligence, intelligence, C. Constantin Poindexter, counterespionage;

The two-hundred-month federal sentence imposed on U.S. Navy sailor Jinchao Wei, also known as Patrick Wei, is not merely a cautionary tale about a single insider’s betrayal. It is a contemporary, well documented case study in the People’s Republic of China’s persistent espionage campaign against U.S. defense entities, executed through an operational pattern that has become all too familiar to counterintelligence practitioners, i.e., low friction spotting and assessment via online platforms, cultivation under plausible non-official cover, incremental tasking that begins with seemingly innocuous collection, and compensation methods that leave a financial signature even when communications are migrated to encrypted channels (U.S. Department of Justice, 2023; U.S. Department of Justice, 2026a). The Wei matter is also a reminder that insider threats rarely begin with the theft of a crown jewel. They begin with ego, attention, a sense of being chosen, and the seductive illusion that the handler is impressed and that the target is smarter than the system.

Public reporting and Department of Justice releases describe Wei as having been arrested in August 2023 as he arrived for duty at Naval Base San Diego, where he was assigned to the amphibious assault ship USS Essex (U.S. Department of Justice, 2023; U.S. Department of Justice, 2026b). The arrest timing and location are operationally significant. Counterintelligence cases often culminate when investigators can control the environment, secure devices and storage, and prevent further loss of national defense information while preserving evidentiary integrity. The government’s narrative, as presented publicly, reflects a mature, documentable case anchored in communications and payment records rather than exotic or highly sensitive sources. The Department of Justice has been explicit that not every investigative step can be disclosed and I don’t intend to do so here, but it has been equally clear that the evidentiary core included intercepts of communication between Wei and his PRC handler, and documentation of how Wei was rewarded for his betrayal (U.S. Department of Justice, 2026a).

The recruitment vector in this case aligns with PRC modus operandi in insider targeting. Wei was approached through social media by an individual presenting as a “naval enthusiast” who claimed a connection to China’s state-owned shipbuilding sector, a cover story designed to appear adjacent to legitimate maritime interest while still close enough to naval affairs to justify pointed questions (U.S. Department of Justice, 2026a; Associated Press, 2026). That presentation is instructive. It reduces the psychological barrier to engagement, provides a rationale for curiosity-driven dialogue, and permits gradual escalation from general discussion to tasking. A handler does not need immediate access to classified networks to create damage. He needs a human source who can provide operationally relevant details, and then he needs to keep the source talking long enough to normalize betrayal.

Once engaged, Wei’s operational security behavior demonstrates both awareness and complicity. He told a Navy friend that the activity looked “quite obviously” like espionage and, after that realization, he shifted communications to a different encrypted messaging application that he believed was more secure (U.S. Department of Justice, 2026a; USNI News, 2026). This is an important marker for investigators and security managers. When a cleared person acknowledges illicit intent yet continues, the motivation is not confusion. It is volition. The move to a “more secure” platform is also characteristic of PRC handling in HUMINT collection. Chinese FIS does not need to provide sophisticated technical tradecraft if the target will self-generate it. Public charging language indicates agreed steps to conceal the relationship, including deletion of conversation records and use of encrypted methods, which reflects basic but purposeful counter-surveillance and denial behavior (U.S. Department of Justice, 2023).

Tasking, as described in public releases, combined opportunistic collection with specific collection requirements. Wei was asked to “walk the pier” and report which ships were present, provide ship locations, and transmit photos and videos along with ship-related details (U.S. Department of Justice, 2026a). From a counterintelligence perspective, these are not trivial asks. Pier-side observations can support pattern of life analysis, readiness inference, and operational planning, particularly when fused with open source material and other clandestine reporting. The case officer’s methodology is “incrementalism”. A handler begins with items that feel observational and deniable, then pulls the source toward more sensitive materials by normalizing the exchange relationship and introducing compensation.

The most damaging element is the alleged transfer of classified technical and operational documents. DOJ accounts state that over an approximately 18-month relationship, Wei provided approximately sixty manuals and other sensitive materials, including at least thirty manuals transmitted in one tranche in June 2022, some of which clearly bore export control warnings. The materials were related to ship systems such as power, steering, weapons control, elevators, and damage and casualty controls (U.S. Department of Justice, 2026a; U.S. Department of Justice, 2026b; Associated Press, 2026). In counterintelligence risk terms, technical manuals provide adversaries with a low-cost blueprint for exploitation. They can inform electronic attack planning, maintenance and sustainment targeting, and vulnerability discovery. They also enable synthetic training and doctrine development for adversary operators. A single manual can be operationally relevant for years because systems and procedures often evolve incrementally, not continuously.

Compensation details illuminate tradecraft and investigative leverage. Wei received more than $12,000 over the course of the relationship, including an alleged $5,000 payment connected to the June 2022 manual transfer. The DOJ has described the use of online payment methods (U.S. Department of Justice, 2023; U.S. Department of Justice, 2026a). This is common in modern espionage involving HUMINT assets who are not professional intelligence officers. Financial transfers create documentary evidence, establish quid pro quo, and provide prosecutors with a corroborating narrative that is legible to a jury. For counterintelligence professionals, this observation is instructive. When communications shift to encrypted platforms, payment flows often remain discoverable through records, device artifacts, and third-party reporting. The operational discipline required to truly eliminate financial signatures is rarely present in an insider unless he or she is COMSEC sophisticated.

Public disclosures describe the case’s investigative architecture in broad but meaningful terms which are instructive even in the absence of the classified story. The FBI and Naval Criminal Investigative Service conducted the investigation. The DOJ characterized the matter as a “first of its kind” espionage investigation in the district, language that signals a substantial investigative effort and a prosecutorial commitment to proving the national security dimension in open court (U.S. Department of Justice, 2026a). The described evidence set emphasizes calls and electronic and audio messages with the PRC handler, payment records and receipts, and a post-arrest interrogation in which Wei admitted to providing the materials and described his conduct as espionage (U.S. Department of Justice, 2026a). Those elements are not glamorous, but they are decisive. They reflect the fundamentals of counterintelligence case building: document the relationship, document tasking and exchanges, document intent and benefit.

This IS PRC modus operandi! The Wei case fits a familiar pattern. The approach was enabled by digital access to targets, the cover identity was plausibly adjacent to the target’s professional interests, and the relationship was escalated through a play on Wei’s ego, . . . a mix of attention, manipulation, and money to compromise him. Tradecraft relied on human psychology, not advanced technical means. The Chinese FIS officer did not need to defeat a classified network. He convinced an insider to carry information out through routine channels and to do so voluntarily. This is a good example of why insider threat programs cannot focus only on clearance adjudication and periodic training. They must incorporate behavioral indicators, targeted education about online elicitation, and strong reporting pathways that reward early disclosure rather than stigmatize it (U.S. Department of Justice, 2023; U.S. Department of Justice, 2026a).

There is also a supervisory and cultural lesson embedded here. Wei voiced suspicion to another sailor. That disclosure was a moment when the damage could have been immediately contained. Peers often see the first signs of a peril, yet peers hesitate, either because they do not want to “ruin someone’s career” or because they assume someone else will act. Counterintelligence operators should treat this as a design requirement. Reporting must be made psychologically easy, procedurally simple, and institutionally supported. A peer report should trigger a calibrated and coordinated response, not an immediate public spectacle. The goal is to get ahead of compromise, not to create an environment where personnel conceal concerns to avoid attention.

The Wei case is a well-evidenced illustration of PRC espionage tradecraft against the United States. Chinese FIS spots and contacts potential insiders at scale through social platforms, cultivates via plausible identity, normalizes secret communications, introduces tasking that begins with the innocuous then escalates to classified materials, and pays through channels that are convenient to the target while still supporting handler control and a firm compromise of the asset (U.S. Department of Justice, 2023; U.S. Department of Justice, 2026a; USNI News, 2026). In my professional judgment, this is another textbook example of ego as the primary driver beneath the surface rationalizations. Even when loneliness, financial temptation, or grievance are present, the consistent psychological engine in treasonous espionage is the ego’s appetite to feel important, chosen, liked, befriended and exceptional. Wei’s conduct underscores that dynamic. He recognized the espionage for what it was, believed he could manage his exposure by encrypted applications, and continued down the road of betrayal. That is not naïveté. It is a belief that rules apply to others, that risk can be controlled by personal cleverness, and that the handler’s attention is a validation of one’s importance in the world. In very few espionage cases, money is the hook. The I.C. likes to think that examples like the Ames Case was a money-motivated treason. It was only partially. Likewise, the I.C. report on Ana Montés lays the blame at the feet of “ideology”. That really wasn’t it. Ego is the line that keeps the source from walking away when conscience and common sense offer an exit. It is almost ALWAYS ego.

~ C. Constantin Poindexter, MA in Intelligence, Graduate Certificate in Counterintelligence, JD, CISA/NCISS OSINT certification, DoD/DoS BFFOC Certification

Bibliography

  • Associated Press. (2026, January 12). Former Navy sailor sentenced to 16 years for selling information about ships to Chinese intelligence.
  • U.S. Department of Justice. (2023, August 3). Two U.S. Navy servicemembers arrested for transmitting military information to the People’s Republic of China.
  • U.S. Department of Justice. (2026a, January 13). Former U.S. Navy sailor sentenced to 200 months for spying for China.
  • U.S. Department of Justice. (2026b, January 14). U.S. Navy sailor sentenced to more than 16 years for spying for China.
  • USNI News. (2026, January 13). Sailor to serve 16 year prison sentence for selling secrets to China.
Posted on

When Counterintelligence Did Not “Catch” Jonathan Soong

espionage, counterespionage, intelligence, counterintelligence, spy, spies, C. Constantin Poindexter

When Counterintelligence Did Not “Catch” the Bad Guy: How Export Compliance and Oversight Stopped an Illicit Transfer

As a counterintelligence guy, I would love to claim one for the team, telling you a story of how counterintelligence “caught” Jonathan Soong. The question presumes a familiar arc: a clandestine plot detected by a vigilant counterintelligence service, followed by an investigative takedown. In practice, many of the most consequential national security cases in the defense industrial base begin elsewhere. They begin in the unglamorous terrain of export controls, contractual oversight, documentation requirements, and compliance escalation. The Soong matter is best read not as a story of counterintelligence brilliance at the point of origin, but as a demonstration that a robust compliance mechanism can function as a practical counterintelligence force multiplier, surfacing deception through audit friction, verification, and internal accountability (U.S. Department of Justice 2025a).

Jonathan Yet Wing Soong worked under a University Space Research Association arrangement supporting NASA, where he helped administer licensing and distribution of U.S. Army-owned aviation and flight control software subject to U.S. export controls. Public charging and plea materials describe a pattern that is familiar to any counterintelligence professional who has studied insider-enabled technology transfer. A trusted administrator leveraged authorized access to facilitate improper export to a prohibited end user, while using misrepresentation and intermediaries to reduce detection risk and sustain the activity long enough to monetize it (U.S. Department of Justice 2022; U.S. Department of Justice 2023; U.S. Department of Commerce, Bureau of Industry and Security 2022).

Export compliance as counterintelligence by another name

In the contractor ecosystem, counterintelligence is no longer confined to investigations and briefings. It is built into controls that regulate who can access what, who can receive what, and what documentation must exist to justify a transfer. Export compliance is the legal expression of strategic technology denial. When an export compliance program is mature, it creates a perimeter of verification around controlled software, technical data, and sensitive know-how. It does this through end-user screening, licensing checks, record retention, and the expectation that representations are auditable, not merely asserted (U.S. Department of Justice 2025a).

Soong’s conduct, as publicly described, involved providing controlled U.S. Army aviation software to the Beijing University of Aeronautics and Astronautics, commonly known as Beihang University, an end-user on the U.S. Entity List. The Entity List designation matters because it transforms what might otherwise be a complicated compliance decision into a bright-line restriction: an elevated risk recipient that generally requires licensing and heightened scrutiny. In counterintelligence terms, it is a government signal that a recipient is associated with activities of concern and therefore must be treated as a strategic risk, not just a commercial counterparty (U.S. Department of Commerce, Bureau of Industry and Security 2022; U.S. Department of Justice 2022).

The decisive tripwire was oversight, not classic counterintelligence detection

The core point that the public often misses is timing. The publicly documented narrative indicates that the scheme was not halted because counterintelligence detected hostile tasking in real time. Rather, the activity began to unravel when NASA asked questions about software licensing activity involving China-based purchasers. That inquiry triggered internal examination at USRA, which then forced Soong’s process, documentation, and representations into a higher scrutiny environment (U.S. Department of Justice 2025a).

From a former operator’s perspective, that is the moment the system displayed its value. Oversight created heat. Heat compelled review. Review compelled proof. Proof created contradictions. Contradictions produced admissions and preserved evidence. That sequence is not incidental. It is the operational logic of compliance as an investigative engine. When a compliance system is designed to verify rather than merely record, it becomes difficult for an insider to sustain a cover story indefinitely.

The cover story failed under verification pressure

Public DOJ descriptions emphasize that Soong initially lied and fabricated evidence to make it appear that purchaser diligence had been conducted. In my experience, this is the most common failure mode for organizations that treat compliance as a box-checking function: insiders learn the minimum artifacts that satisfy superficial review. The Soong case illustrates what happens when counsel and compliance do not accept the first answer. DOJ accounts describe further investigation by USRA’s counsel, confrontation with contradictions, and Soong’s eventual admissions, including that he knew the end user was on the Entity List and that an export license was required (U.S. Department of Justice 2025a).

That is not just a legal detail. It is the fulcrum that turns suspicion into provable intent. Counterintelligence professionals care about intent because intent distinguishes mistake from exploitation and distinguishes weak governance from an insider who is actively enabling a strategic competitor or worse, adversarial FIS. Admissions anchored to documented contradictions are highly durable. They are not dependent on classified sources or contested analytic judgments. They are built for court cases.

Intermediaries and misdirection are a compliance evasion pattern

The public record also describes the use of an intermediary to obscure the true end user and facilitate the commercial pathway. This is a standard concealment vector. Intermediaries can be used to launder payment trails, shift transactional geography, and create plausible deniability within internal processes that rely on surface-level end-user statements. If a program relies on the integrity of a single administrator’s “screening,” the administrator becomes the control. If the administrator is compromised, the system is compromised. In this case, public materials describe intermediary involvement and a transfer pathway that, when examined, revealed the underlying restricted recipient (Department of Defense Office of Inspector General, Defense Criminal Investigative Service 2023; U.S. Department of Justice 2025a).

For counterintelligence practitioners, the lesson is straightforward: third party structures are not merely procurement conveniences. They are also tradecraft. In an export controls environment, every intermediary should be treated as a potential concealment method unless diligence is independently verifiable.

Voluntary self-disclosure converted an internal discovery into a national security case

Once internal discovery occurred, the matter moved from corporate governance to national security enforcement. DOJ’s public declination notice emphasized that USRA self disclosed export control offenses committed by its employee and cooperated, which shaped the government’s posture toward the company while leaving the individual to face prosecution (U.S. Department of Justice 2025a). That sequence is important for practitioners because it demonstrates how compliance maturity affects outcomes. Prompt internal escalation, self disclosure, and remediation can separate an organization’s institutional exposure from the conduct of a rogue insider, while also strengthening the government’s ability to build a case against the perpetrator.

DOJ also identified the investigative constellation, including Commerce export enforcement, the FBI, Defense Criminal Investigative Service, NASA Office of Inspector General, and U.S. Army elements including Army counterintelligence and investigative components. In other words, counterintelligence was present and relevant, but it was not the initial tripwire. It was part of the enforcement and investigative consolidation phase after compliance mechanisms surfaced the issue and the company disclosed it (U.S. Department of Justice 2025a; U.S. Department of Justice 2023).

Compliance “caught” the act and counterintelligence helped finish the job

If we insist on the verb “catch,” my professional assessment is that counterintelligence did not “catch” Jonathan Soong in the popular sense of the term. The decisive early detection function was performed by oversight and export compliance mechanisms. NASA’s questions triggered organizational scrutiny. Scrutiny demanded documentation. Documentation collapsed under verification. Verification produced contradictions and admissions. Those admissions and records enabled self-disclosure and a multi-agency investigation that culminated in a guilty plea. Counterintelligence contributed where it often contributes most effectively in the contractor environment: by supporting the investigative and enforcement architecture once a compliance tripwire has surfaced misconduct, and by helping translate a technical compliance failure into a national security narrative that the government can prosecute (U.S. Department of Justice 2025a; U.S. Department of Justice 2023).

This is not a criticism of counterintelligence. It is an argument for modernizing how we describe counterintelligence effectiveness. In the defense industrial base, export compliance is not adjacent to counterintelligence. Export compliance is frequently counterintelligence in operational form. When built correctly, it makes illicit transfer hard to hide, expensive to sustain, and likely to fail under audit pressure. The Soong case is the quiet proof that governance, oversight, and export controls can stop a technology transfer plot even when no one is running a classic counterintelligence operation at the beginning.

~ C. Constantin Poindexter, MA in Intelligence, Graduate Certificate in Counterintelligence, JD, CISA/NCISS OSINT certification, DoD/DoS BFFOC Certification

Bibliography

  • Department of Defense Office of Inspector General, Defense Criminal Investigative Service. 2023. “Defendant Admits Using Intermediary to Funnel Payments for United States Army Aviation Software Exported to Beihang University.” Press release, January 17, 2023.
  • U.S. Department of Commerce, Bureau of Industry and Security. 2022. “South Bay Resident Charged with Smuggling and Exporting American Aviation Technology to Beijing University.” Press release, May 26, 2022.
  • U.S. Department of Justice. 2022. “South Bay Resident Charged with Smuggling and Exporting American Aviation Technology to Beijing University.” Press release, U.S. Attorney’s Office, Northern District of California, May 26, 2022.
  • U.S. Department of Justice. 2023. “Castro Valley Resident Pleads Guilty to Illegally Exporting American Aviation Technology.” Press release, U.S. Attorney’s Office, Northern District of California, January 17, 2023.
  • U.S. Department of Justice. 2025a. “Justice Department Declines Prosecution of Company That Self Disclosed Export Control Offenses Committed by Employee.” Press release, Office of Public Affairs, April 30, 2025.

Posted on

The Power of OSINT: Attribution and the Identification of Oleg Smolenkov

spy, spies, intelligence, counterintelligence, counterespionage, C. Constantin Poindexter Salcedo, national security, NATSEC

The public identification of Oleg Smolenkov illustrates a contemporary counterintelligence reality, i.e., a sensitive resettled asset can be unmasked without a hostile service penetrating classified systems, without a HUMINT penetration of our services, and without a single administrative branch leak. The decisive mechanism is open-source convergence, the disciplined fusion of administrative traces, archived reporting, and behavioral cues into a coherent attribution. In this case, a non-governmental investigative newsroom and parallel OSINT collectors did not begin with a name. They began with a publicly reported event, a bounded time window, and an implied access profile. They exploited predictable transparency mechanisms in both Russia and the United States. The result was a high confidence linkage between an anonymized description of a relocated source and a specific individual, supported by cross-domain corroboration. (Bellingcat 2019; Reuters 2019a)

A counterintelligence practitioner evaluating this episode should resist the temptation to treat it as an exceptional scandal driven by personalities or politics. It is better understood as a repeatable analytic pipeline. The steps are familiar: cueing, candidate generation, plausibility testing, registry linkage, reaction validation, and signature reinforcement. Each step relies on data that appears mundane in isolation. The compromise emerges from aggregation.

The process begins with cueing. In early September 2019, major United States media described a clandestine extraction that occurred in 2017 involving a high-level Russian government source who had provided unusually sensitive insight into Kremlin and Russian policy-level decision-making. The reporting framed the extraction as protective and urgent, occurring after fears that the source’s security was at risk. Even when anonymized, those details are operationally useful to a determined investigator because they narrow the search space. The investigator obtains institutional scope, timing, and a risk narrative, which in counterintelligence terms function as selection criteria. The relevant question becomes: which Russian official with plausible access disappears from public view in the relevant period, under circumstances consistent with sudden relocation? (Time 2019; RFE RL 2019b)

Once cueing is in place, candidate generation becomes feasible. OSINT investigators queried Russian language media archives, cached pages, and secondary reporting for disappearance stories in the mid-2017 window that involved government personnel. The case benefited from pre-existing Russian reporting. Multiple outlets later described that Russian authorities had opened a criminal investigation in 2017 into the suspected murder of a missing official who disappeared during travel in Montenegro, and that the investigation was later abandoned after authorities concluded the individual was alive abroad. The Guardian reported that the online outlet Daily Storm had described that arc, including the murder probe and the eventual conclusion that the official had left Russia. (The Guardian 2019a; RFE RL 2019c) Reuters likewise reported that Kommersant identified the possible individual as Oleg Smolenkov, describing his disappearance in Montenegro in June 2017 with his wife and children, along with the evolution of the Russian investigative posture from suspected murder to an assessment that he was living abroad. (Reuters 2019a)

The OSINT collector has a candidate name and an event narrative that already fits the timing constraint. A counterintelligence practitioner will note the structural weakness revealed here. If an extraction or relocation corresponds to a conspicuous real-world absence, and if that absence triggers a foreign law enforcement process, then the foreign process itself can generate discoverable artifacts, including press interest, investigative leaks, and later retrospective reporting. Even if the foreign process is opaque, the fact pattern is often newsworthy enough to be recorded ‘somewhere’, and later rediscovered when a cueing event directs attention to it.

We now move to plausibility testing. A candidate must match the access and placement implied by the original extraction narrative. Collectors therefore, reconstruct a career trace from open sources. Reuters reported that the Kremlin confirmed that a person named Smolenkov had worked in the Russian presidential administration and had been dismissed, while disputing that he had meaningful access to President Vladimir Putin. (Reuters 2019a; Reuters 2019b) Whether or not one accepts the Kremlin’s minimization, the acknowledgement of employment is itself confirmatory for attribution purposes. This validates that the named candidate is not fictional, and places the asset inside the relevant institutional universe.

Supplementary open source synthesis connected Smolenkov to senior foreign policy structures, particularly through reporting that he had worked in the Russian embassy in Washington during a period associated with senior diplomat Yuri Ushakov and later served in roles linked to the presidential administration. Russia Matters summarized Kommersant reporting that described Smolenkov as a longtime assistant to Ushakov, which is precisely the kind of staff proximity that can produce indirect exposure to high-level deliberations without public prominence. (Russia Matters 2019) From a counterintelligence perspective, that distinction matters. A source does not need to be a cabinet-level decision maker to be strategically valuable. In many systems, staff, aides, and administrators are the connective tissue that accesses documents, schedules, and briefing flows. OSINT collectors correctly treat that staff layer as a plausible access vector.

Plausibility testing alone still does not establish that the candidate is the person relocated to the United States. The decisive linkage emerged from United States administrative records, particularly property ownership documentation. Bellingcat reported that open records showed “Oleg and Antonina Smolenkov” purchasing a home in northern Virginia in June 2018, and connected that purchase to the hypothesis that the family had been resettled under protection after leaving Russia. (Bellingcat 2019) RFE RL reporting similarly discussed public records indicating ownership of a house in Stafford County, Virginia, by Oleg and Antonina Smolenkov, and described subsequent changes consistent with attempts to reduce visibility, including the transfer of ownership into a trust. (RFE RL 2019a; RFE RL 2019b)

For a counterintelligence practitioner, this phase is the core operational lesson. The United States property recording system is designed to be durable, searchable, and transparent. A relocated human source living under a real name, or under a name that can be linked by deed chain, becomes discoverable. Even when a trust is used, the initial purchase may preserve the identity in a durable record, and later transfers can be traced. The trust can help against casual discovery, but it does not reliably defeat an investigator who already has a lead and is willing to follow the chain across databases. Even to the untrained eye, recent deeding from a Russian surname to a blind trust is a dead giveaway.

The deed stuff is important, HOWEVER, the linkage was not limited to property records in isolation. Investigators layered temporal correlation. The property purchase followed the 2017 disappearance window by roughly one year, a plausible period for relocation, debriefing, and resettlement logistics. The geographic placement, near Washington, aligns with the practical needs of ongoing handling, access to government liaison, and security support. RFE RL reporting placed the residence in a neighborhood with current and former United States government personnel, which would not be an implausible environment for a protected relocatee, but also increases the risk of attention because residents recognize unusual patterns. (RFE RL 2019a)

Attribution confidence increases through cross-side corroboration. Reuters reported that Russian state media and other Russian outlets visited or referenced the Virginia address associated with Smolenkov, and that Russian official commentary focused on disputing his access level rather than disputing his identity. (Reuters 2019b) The Guardian reported that Russian media quickly identified Smolenkov as the likely figure after the initial extraction story circulated and that earlier Russian reporting had already treated his disappearance as suspicious. (The Guardian 2019b) In counterintelligence analytic terms, this is validation by reaction. When an implicated government acknowledges employment, debates seniority, and frames narratives around access, it implicitly accepts the identity anchor, even if it contests the operational characterization.

Another reinforcing layer is signature observation, i.e., the detection of behaviors consistent with protective posturing. Bellingcat described journalists encountering indications of security presence when approaching the residence and noted that the family likely departed after the story circulated. (Bellingcat 2019) RFE RL reported that neighbors stated that the family of the identified property left abruptly soon after publicity, and that no one answered at the residence when a reporter from RFE visited. (RFE RL 2019a; RFE RL 2019b) From a practitioner’s view, these signatures are a bit ambiguous but directionally meaningful. They do not prove intelligence affiliation however, they do add coherence to the broader narrative when combined with verified administrative linkages.

The identification of Smolenkov can be described as an open-source attribution chain with mutually reinforcing elements. The chain begins with an anonymized description of an extracted asset, which supplies a time-bound and an access profile. It then leverages a pre-existing disappearance narrative in Russian reporting that matches the window. It validates institutional plausibility through official acknowledgement of employment and through open source reconstruction of staff level proximity to senior policy structures. It then bridges the gap from Russia to the United States by locating the same names in property records, supported by temporal correlation and geographic plausibility. Finally, attribution through adversary reaction and observable protective signatures after publicity adds stability to the former. Each element could be dismissed alone, but in totality of the circumstances they provide a high probability attribution that is operationally sufficient for pretty damn reliable public identification. (Bellingcat 2019; Reuters 2019a; RFE RL 2019b)

The counterintelligence implications are clear. A protective extraction does not end an operational dilemma. It begins a new phase in which the threat is not hostile surveillance alone but also open-source exploitation. Transparency regimes create predictable exposure surfaces. Registers of deeds and county recorders, tax collectors, court records, licensing agencies, and corporate filing records are not intelligence sources, but they are an extremely searchable source of structured and more than reasonably accurate data. C.I. measures or countermeasures applied after asset resettlement (such as transferring property into a trust) reduce opportunistic discovery after the fact but will fail against an OSINT collector that already possesses a starting point. Also, cueing can be powerfully exploited. Public narratives about timing and sensitivity can provide sufficient structure for a collector to find pre-existing anomalies and connect them to these domestic records.

The Smolenkov resettlement demonstrates that our source protection doctrine must be extended beyond traditional clandestine concerns. It must incorporate administrative footprint management, name and identity compartmentation, and a realistic appreciation of how quickly digital records can be correlated across jurisdictions, in real-time and remotely. This case CANNOT be viewed as an isolated breach. It really is a warning about the baseline capabilities of OSINT. Oh, and a parting shot from a former C.I. guy, don’t put properties that are deeded to Russian surnames into trusts that are filed publicly.

~ C. Constantin Poindexter, MA in Intelligence, Graduate Certificate in Counterintelligence, JD, CISA/NCISS OSINT certification, DoD/DoS BFFOC Certification

Bibliography

  • Bellingcat. 2019. “Murdered in Montenegro, or Living in Suburban Virginia? Unraveling the 2017 American Spy Story.” September 10, 2019.
  • Radio Free Europe Radio Liberty. 2019a. “Virginia Residents Question Whether Their Neighbor Was a Russian Informant.” September 10, 2019.
  • Radio Free Europe Radio Liberty. 2019b. “Russia Seeking Interpol’s Help on Location of Alleged CIA Informant.” September 12, 2019.
  • Radio Free Europe Radio Liberty. 2019c. “Paper Reports ‘Details’ of Alleged CIA Informer’s Disappearance in Montenegro.” September 12, 2019.
  • Reuters. 2019a. “Kremlin Says Alleged U.S. Spy Did Not Have Access to Putin.” September 10, 2019.
  • Reuters. 2019b. “Russia Blasts Idea a CIA Mole Lifted Lid on Its U.S. Meddling.” September 11, 2019.
  • Russia Matters. 2019. “Russia in Review, Sept. 6 to 13, 2019.” September 2019.
  • The Guardian. 2019a. “Russia Investigated Disappearance of Suspected US Spy as Possible Murder.” September 10, 2019.
  • The Guardian. 2019b. “Oleg Smolenkov: Alleged US Spy Who Gave Russia the Slip.” September 14, 2019.
  • Time. 2019. “The U.S. Reportedly Extracted a High Level Spy From Russia in 2017 Amid Concerns of Mishandled Intelligence.” September 10, 2019.
Posted on

Conduct Not Becoming: Alleged U.S.-Linked Interference in Greenland

intelligence, counterintelligence, espionage, counterespionage, spy, spies, subversion operations, c. constantin poindexter

In August 2025, the Danish Ministry of Foreign Affairs summoned the chargé d’affaires of the United States Embassy in Copenhagen after revelations by the Danish Broadcasting Corporation (DR) that several Americans linked to the U.S. Presidential Administration had engaged in covert political activities in Greenland. According to the reporting, these individuals compiled lists of Greenlanders categorized as “pro-U.S.” or “anti-Trump,” cultivated ties with local elites, and promoted narratives designed to widen divisions between Nuuk and Copenhagen (PBS NewsHour 2025; Associated Press 2025). The Danish Security and Intelligence Service (PET) emphasized that Greenland remains a primary target for foreign interference operations, echoing earlier warnings that external actors could exploit or fabricate political disagreements within the Kingdom of Denmark (Al Jazeera 2025). These revelations, coupled with Denmark’s unusually direct diplomatic response, illustrate the characteristics of a malign influence or subversive operation and highlight the potential damage such activities can inflict on U.S. national security, particularly by undermining liaison trust with Denmark, arguably one of Washington’s most important intelligence partners.

Characteristics of a Malign Influence Operation

Malign influence operations are typically defined by certain recurring attributes: plausible deniability, use of cut-outs or intermediaries, audience mapping and segmentation, amplification of divisive narratives, and efforts aimed at shaping decision-making environments rather than openly persuading through argument. The Greenland episode, as described by DR and reported internationally, bears all of these hallmarks.

Plausible deniability was central to the actor’s positioning. Officials stressed that the government does not direct or control the actions of private citizens, even though the actors were reportedly politically connected to the Administration (PBS NewsHour 2025). Such disavowals allow states to shield themselves from direct accountability while benefiting from the effects of covert activity.

The use of cut-outs and informal networks appears evident. The alleged operatives were not formal embassy staff operating under Chief of Mission authority but instead American nationals cultivating relationships with Greenlandic political and business figures. This indirect approach mirrors tradecraft seen in both Cold War–era and contemporary influence campaigns, allowing sponsors to maintain distance while pursuing strategic objectives (Associated Press 2025).

The activity involved audience segmentation, as evidenced by the preparation of lists distinguishing sympathetic Greenlanders from opponents. Such mapping is a well-established precursor to micro-targeted persuasion and coalition-building (Rudbeck 2020).

The operation sought to exploit existing grievances. Greenland has a long history of tension regarding its relationship with Copenhagen, particularly concerning autonomy and resource management. PET has publicly warned that adversaries attempt to “promote or amplify particular viewpoints” in Greenland to exacerbate these tensions (Al Jazeera 2025). By pressing sovereignty-oriented narratives, the actors aligned with known fault lines.

The activities pursued a strategic effect on governance: nudging Greenland’s politics toward greater separation from Denmark or, at minimum, intensifying friction between Copenhagen and Nuuk. This fits the definition of a malign influence campaign, which aims not merely to influence public opinion but to shift the constitutional or strategic environment of the target polity (Fleshman et al. 2020).

Greenland’s Strategic Importance

Understanding why Denmark reacted so firmly requires recognition of Greenland’s immense strategic value. The island hosts Pituffik Space Base (formerly Thule Air Base), the northernmost U.S. military installation. Pituffik is critical to ballistic missile early warning, missile defense, and space surveillance missions, particularly through the 12th Space Warning Squadron, which tracks ballistic launches and supports U.S. Space Force operations (U.S. Space Force 2024). Pituffik’s radar and space-tracking systems are a vital component of NATO deterrence, as they enable early detection of potential Russian or other adversary launches.

Beyond Pituffik, Greenland’s geography makes it indispensable to North Atlantic security. The island sits astride the Greenland–Iceland–U.K. (GIUK) gap, a maritime chokepoint central to monitoring Russian submarine traffic from the Barents Sea into the North Atlantic. As the Arctic becomes increasingly contested due to climate change and resource competition, Greenland’s location at the intersection of North America and Europe magnifies its strategic importance.

Equally significant is Denmark’s intelligence capability. The Danish Defence Intelligence Service (FE) and PET are widely regarded as among the most capable small-state services, particularly in signals intelligence, counterintelligence, cyber and Arctic domain awareness. FE’s Intelligence Risk Assessment 2024 explicitly identifies Greenland as a frontline in great-power competition (Danish Defence Intelligence Service 2024). As such, Denmark is one of Washington’s most important intelligence liaisons, and trust in this relationship is crucial to U.S. and NATO security.

Damage to U.S. National Security

From a U.S. perspective, even if the federal government neither authorized nor directed the actions of the Americans involved, the perception of interference inflicts real costs. Four national security risks stand out.

Such activities risk eroding liaison trust. Intelligence sharing relies on reciprocity and respect for sovereignty. If Denmark perceives that the United States tolerates or encourages efforts to manipulate the Kingdom’s internal affairs, Danish services may hesitate to share sensitive information or to cooperate fully in Arctic monitoring. Trust, once diminished, is difficult to rebuild (Danish Defence Intelligence Service 2024).

Malign influence in Greenland undermines coordinated Arctic policy. Pituffik’s continued operation depends on alignment among Copenhagen, Nuuk, and Washington. Any perception that the U.S. is fueling secessionist sentiment in Greenland complicates base access negotiations, environmental approvals, and trilateral defense arrangements. Diplomatic friction could translate into delays or restrictions that weaken early warning and space-tracking capabilities (Associated Press 2025).

Such revelations hand adversaries narrative ammunition. Russia and China have long sought to depict the United States as hypocritical in its advocacy for sovereignty and democratic norms. A Danish finding that U.S.-linked actors engaged in subversive activity in Greenland would provide propaganda fodder for Moscow and Beijing, undermining U.S. credibility in NATO and in multilateral Arctic governance forums (Al Jazeera 2025).

These operations jeopardize operational continuity at Pituffik. Strategic deterrence depends on uninterrupted coverage of missile warning and space tracking. Political discord that affects budgets, labor relations, or local sentiment in Greenland could generate friction costs that weaken U.S. posture in the High North (U.S. Space Force 2024).

Theoretical Framing: Rudbeck and Malign Influence

Emma Rudbeck’s (2020) master’s thesis on foreign interference in Greenland provides an instructive framework. Applying Applied History and strategic narrative theory, Rudbeck argues that interference by major powers in Greenland echoes Cold War–era dynamics and threatens the concept of “Arctic Exceptionalism,” which had long portrayed the region as insulated from great-power rivalry. She concludes that Denmark must prepare for sustained interference by China, Russia, and the United States, and recommends a proactive Arctic strategy that emphasizes resilience and narrative management. Rudbeck’s insights align with the Greenland episode. The use of covert actors to segment populations and inflame tensions fits her description of “strategic narratives” designed to reshape perceptions of sovereignty and autonomy. By treating Greenland not as a neutral space but as contested political terrain, the alleged U.S.-linked operatives validated Rudbeck’s claim that interference is no longer limited to Russia or China but includes Washington itself. From Denmark’s perspective, this raises uncomfortable questions about the reliability of its closest ally.

Assessing the “Deception Operation” Frame

Denmark’s choice to summon the U.S. envoy demonstrates that it viewed the incident not as isolated private advocacy but as a coherent deception operation. The tactics of covert list-building, elite cultivation, and narrative seeding abroad reflect classic subversive tradecraft, intended to give the false impression of grassroots political momentum. PET has warned precisely about such techniques, noting that foreign influence in Greenland often seeks to “amplify particular viewpoints” to sow division (Al Jazeera 2025). This aligns with broader theoretical work on deception and influence, which emphasizes how adversaries shape decision environments by hiding their involvement (Fleshman et al. 2020).

We Need to Assure Copenhagen that “This isn’t who we are”

The Greenland case illustrates how malign influence can damage alliances even when conducted by non-official actors. Mitigating this damage will require visible U.S. steps: clear ministerial-level assurances to Copenhagen, tighter deconfliction to ensure all outreach in Greenland is coordinated through embassy channels, and symbolic trilateral initiatives with Denmark and Greenland to demonstrate respect for the Kingdom’s internal constitutional order. Absent such efforts, suspicion of U.S. duplicity may persist, weakening NATO cohesion at a time when Arctic security is increasingly central.

The alleged Administration-linked interference in Greenland demonstrates the characteristics of a malign influence operation: plausible deniability, cut-outs, audience segmentation, exploitation of grievances, and pursuit of strategic effects on governance. Greenland’s unique importance to U.S. defense posture and Denmark’s role as an elite intelligence ally magnify the stakes. By alienating Copenhagen, such operations risk degrading liaison trust, undermining trilateral defense cooperation, handing adversaries propaganda, and jeopardizing early-warning missions at Pituffik. Rudbeck’s (2020) analysis underscores that Greenland is no longer insulated from great-power rivalry, and that even allies may engage in subversive activity. For the United States, this episode should serve as a cautionary reminder that short-term political maneuvering can yield long-term strategic perils, especially when it undermines the trust of one of its most vital partners in the Arctic.

~ C. Constantin Poindexter, M.A. en Inteligencia, Certificado de Posgrado en Contrainteligencia, J.D., certificación CISA/NCISS OSINT, Certificación DoD/DoS BFFOC

References

Al Jazeera. 2025. “Denmark Summons US Envoy over Trump Allies’ Alleged Greenland Interference.” Al Jazeera, August 28, 2025. https://www.aljazeera.com/news/2025/8/28/denmark-summons-us-envoy-over-trump-allies-greenland.

Associated Press. 2025. “Denmark Summons US Envoy over Alleged Trump Allies’ Interference in Greenland.” AP News, August 28, 2025. https://apnews.com/article/denmark-greenland-us-trump-6c9544314792cf1e287e21af06111c1e.

Danish Defence Intelligence Service. 2024. Intelligence Risk Assessment 2024. Copenhagen: FE. https://fe-ddis.dk/en.

Fleshman, William, Jennifer L. Larson, and Christopher Paul. 2020. “Deception and the Strategy of Influence.” arXiv preprint arXiv:2011.01331.

PBS NewsHour. 2025. “Denmark Summons US Envoy over Claims of Interference in Greenland.” PBS NewsHour, August 28, 2025. https://www.pbs.org/newshour/world/denmark-summons-u-s-envoy-over-claims-of-interference-in-greenland.

Rudbeck, Emma. 2020. How Should the Kingdom of Denmark React to the Increased Chinese, Russian, and U.S. Interference in Greenland in Its Coming Arctic Strategy? Master’s thesis, University of Southern Denmark. https://thesis.sdu.dk/download?id=2260.

U.S. Space Force. 2024. “12th Space Warning Squadron, Pituffik Space Base.” Fact Sheet, Department of the Air Force. https://www.spaceforce.mil/About-Us/Fact-Sheets/
.

Posted on

Artificial Intelligence and Offensive Counterintelligence in the U.S. I.C.

counterintelligence, intelligence, espionage, counterespionage, espia, spy, spies, contrainteligencia, contraespionaje, c. constantin poindexter, J2, CNI, DNI

Artificial intelligence is transforming the national security landscape by augmenting the capabilities of intelligence organizations to “identify, disrupt, and neutralize adversarial threats”. While much scholarly and policy attention has been devoted to the defensive applications of AI, i.e., cybersecurity, threat detection, and insider threat monitoring, implications for offensive counterintelligence (CI) are equally profound. Offensive counterintelligence, which involves proactive measures to manipulate, exploit, or dismantle adversarial intelligence operations, has traditionally depended on human ingenuity, deception, and long-term HUMINT. The introduction of AI into this realm promises to exponentially increase the scale, speed, and sophistication of U.S. counterintelligence campaigns. The U.S. Intelligence Community (IC) will become more effective at penetration of FIS, deception operations, and neutralization of espionage activities.

One of the most significant ways AI will enhance offensive counterintelligence is through advanced pattern recognition and anomaly detection across massive data streams. The IC already ingests petabytes of information daily, from open-source intelligence (OSINT) to signals intelligence (SIGINT). Offensive counterintelligence officers have historically been hobbled by fragmentary reports and painfully dry and drawn-out analysis to identify foreign intelligence officers, their networks, and their vulnerabilities. Machine learning algorithms now enable CI analysts to identify subtle anomalies in communications metadata, financial transactions, or travel records that suggest covert operational behavior. Algorithms trained on known espionage tradecraft can detect anomalies in mobile phone usage, repeated travel to consular facilities, or encrypted message timing that would elude traditional analysis (Carter, 2020). By automating the detection of clandestine activity, AI provides offensive CI officers with early targeting leads for recruitment, deception, or disruption.

AI’s role in predictive modeling of adversary behavior is a game-changer. Traditional counterintelligence operations have required years of painstaking collection before a service could anticipate an adversary’s moves. Now, reinforcement learning and predictive analytics can generate probabilistic models of how foreign intelligence services will act under specific conditions. This capability is invaluable for offensive CI, in which anticipating an adversary’s agent recruitment attempts or technical collection strategies and techniques allows the U.S. to insert double agents, conduct controlled leaks, or channel disinformation in ways that compromise foreign intelligence effectiveness (Treverton & Miles, 2021). By simulating adversary decision-making processes and Loops, AI effectively allows the IC to wage a chess match several moves ahead, shifting initiative in favor of U.S. operators.

AI will transform deception operations, a core element of offensive counterintelligence. Deception requires constructing credible false narratives, fabricating convincing documents, and sustaining elaborate covers. Generative AI models provide new tools for producing synthetic but convincing content, i.e., emails, social media profiles, deepfake videos, etc., that can be deployed to manipulate adversarial intelligence targets. These capabilities enable more robust false-flag operations, digital honeypots, and disinformation campaigns designed to lure adversary collectors into traps or consume their resources chasing fabricated leads. Deepfake technology raises concerns about disinformation in democratic societies, however, if deployed in a tightly controlled counterintelligence context it becomes a force multiplier, providing scalable deception tools that previously demanded enormous human and material resources (Brundage et al., 2018).

AI enhances the identification and exploitation of recruitment opportunities, central to offensive CI operations. The IC has long relied on spotting, assessing, and recruiting human assets with access and placement. AI-driven analysis of social media, professional networks, and digital exhaust enables rapid identification of individuals with access, grievances, or vulnerabilities suitable for recruitment. Natural language processing (NLP) tools can detect sentiment, stress, or dissatisfaction in posts, while network analysis maps reveal connections within bureaucracies or security services (Greitens, 2019). By narrowing down large populations to high-value recruitment targets, AI augments human case officer ability to prioritize approaches and customize persuasion angles. The integration of AI with human tradecraft accelerates the traditionally slow and resource-intensive recruitment cycle.

Cyber counterintelligence represents another frontier where AI confers offensive advantages. FISs increasingly operate in cyberspace, exfiltrating sensitive data and conducting influence campaigns. AI-enabled intrusion detection, combined with offensive cyber capabilities, allows U.S. counterintelligence to not only identify intrusions but also manipulate them. AI can facilitate “active defense” strategies in which foreign intelligence hackers are fed false or misleading data, undermining their confidence in purloined data. Automated adversarial machine learning tools can also detect attempts by foreign services to poison U.S. AI training data, allowing counterintelligence operators to preemptively counter them (Henderson, 2022). AI both defends critical systems and creates new opportunities for denial and deception operations (D&D) and disruption of adversarial cyber espionage.

Further, AI also addresses one of the perennial challenges of offensive counterintelligence, scalability. Human operator and analyst resources are finite. Adversarial services often enjoy the advantage of operating from within authoritarian systems unconstrained by meaningful oversight. AI offers the IC the ability to scale counterintelligence operations across global theaters without proportional increases in manpower. Automated triage systems can flag potential espionage indicators for human review, while AI-driven simulations can test the effectiveness of proposed offensive strategies before deployment. The scalability of AI ensures that offensive CI efforts remain proactive rather than reactive, allowing the IC to contest adversarial services at a global level (Allen & Chan, 2017).

I will note here that the insertion of AI into offensive counterintelligence is not a panacea. Overreliance on algorithmic outputs without human validation can lead to “false positives”, misidentification, or ethically and legally problematic targeting. Adversaries are also rapidly adopting AI for their own counter-counterintelligence measures, raising the specter of an AI-driven arms race in deception, espionage and counterespionage disciplines. The U.S. IC must ensure that AI tools are embedded within a robust framework of human review, legal compliance, and ethical norms. Offensive CI, clearly operating in the shadows of democratic accountability, requires enhanced governance mechanisms to balance operational effectiveness with adherence to rule-of-law principles (Zegart, 2022).

The adoption of AI in offensive counterintelligence necessitates organizational adaptation. Case officers, analysts, and technical specialists must be trained not only to use AI tools but also to understand their limitations. Interdisciplinary collaboration between computer scientists, behavioral experts, and intelligence professionals will be essential for designing AI systems that are operationally relevant, a particularly challenging problem in a group of agencies accustomed to “siloing”. Investment in secure, resilient AI infrastructure is critical, as adversaries will inevitably seek to penetrate, manipulate, or sabotage U.S. counterintelligence AI systems. Just as past eras of counterintelligence revolved around protecting codes and agent networks, the new era will hinge on safeguarding the integrity of AI platforms themselves (Carter, 2020).

Artificial intelligence offers unprecedented opportunities to enhance the effectiveness of offensive counterintelligence. By improving anomaly detection, predictive modeling, deception, recruitment targeting, and cyber counterintelligence, AI serves as both a force multiplier and a strategic enabler. It allows the IC to proactively shape the intelligence battlespace, seize the initiative from adversaries, and scale operations to meet global challenges. These opportunities come with risks, ethical, operational, and strategic, however, with careful management the payoff will be monumental. Offensive counterintelligence has always been a contest of wits, deception, and foresight. In the twenty-first century, AI will become the decisive instrument that determines whether the U.S. retains the upper hand in the shadow war.

~ C. Constantin Poindexter, M.A. en Inteligencia, Certificado de Posgrado en Contrainteligencia, J.D., certificación CISA/NCISS OSINT, Certificación DoD/DoS BFFOC

References

Allen, G., & Chan, T. (2017). Artificial intelligence and national security. Belfer Center for Science and International Affairs, Harvard Kennedy School.

Brundage, M., Avin, S., Clark, J., Toner, H., Eckersley, P., Garfinkel, B., … & Amodei, D. (2018). The malicious use of artificial intelligence: Forecasting, prevention, and mitigation. Future of Humanity Institute.

Carter, A. (2020). The future of counterintelligence in the age of artificial intelligence. Center for a New American Security.

Greitens, S. C. (2019). Dealing with demand for authoritarianism: The domestic politics of counterintelligence. International Security, 44(2), 9–47.

Henderson, T. (2022). Offensive cyber counterintelligence: Leveraging AI to deceive adversaries. Journal of Cybersecurity Studies, 8(1), 55–74.

Treverton, G. F., & Miles, R. (2021). Strategic counterintelligence: The case for offensive measures. RAND Corporation.

Zegart, A. (2022). Spies, lies, and algorithms: The history and future of American intelligence. Princeton University Press.

Posted on

The Challenge of Spying on China

spy, spies, espionage, counterespionage, intelligence, counterintelligence,carlyle poindexter, constantin poindexter

The WSJ article on Wednesday (Challenge of Spying on China) is a sad reminder of the United States Intelligence Community’s apparent failure to accomplish any broad covert or clandestine penetration of the People’s Republic of China (PRC) in recent history. The lack of HUMINT human intelligence sources (HUMINT) with meaningful access and placement deprives us of insight into Chinese decision making, immediate strategic threat intelligence and perhaps more importantly, gravely impairs U.S. offensive counterintelligence operations.

Moving beyond the obvious difficulties with HUMINT operations within the PRC, reminiscent of the Cold War hostile operational environments, the Intelligence Community is overdue for a paradigm shift in human asset recruitment methodology. For the better part of the last century, the United States Intelligence Community relied on a steady flow of “walk-ins”, volunteers from opposing foreign intelligence services or governments that offered their countries’ secrets. Intelligence officers enjoyed a large degree of success based on a fairly global perception that Americans were the “good guys”, representatives of the land of fairness, equality and justice, qualities that stood in stark contrast to the ruthless and despotic republics from whence they came. Unfortunately, the mystique has faded leaving outsiders to wonder if the values that we promote to the world are nothing more than a hypocritical farce. Mass diffusion of the “Big Lie” throwing fair elections into question, an attempted coup d’etat by an outgoing president, and military involvement under highly questionable intelligence assessments erode the view once held that the United States is the “shining beacon to the oppressed”.

Chinese citizens enjoy a better standard of living than at any time in China’s history. China can rightfully boast that it is a world power and its population can justifiably be proud of its progress. Personal financial success and pride in country promote loyalty. That there is no broad internal rejection of onerous mass surveillance, social credit controls and ethnic cleansing as is the case with the Uyghurs, is a testament to the PRC’s ability to deny facts, deceive its population and prevent the import of non-PRC approved “truths” about freedom and justice within China. The Chinese cultural tendency to identify with the collective rather than the individual is likewise amplified by the PRC’s massive social control machine, with opposing or antagonistic perspectives effectively blocked by the Great Firewall or simply drowned out of public discourse by the volumes of Party-approved propaganda. The PRC’s strategy has created an environment that is more resistant to traditional intelligence recruitment techniques such as economic coercion, ideology exploitation and ego-stroking. Chinese intelligence service recruiters lean on the cultural affinity of ethnically Chinese living in the United States to turn them into spies, coerce them by alluding to what might become of their families living in China or deploy the time-tested technique of guanxi to achieve intelligence asset recruitments. United States intelligence officers do not enjoy a parallel or equivalent.

FBI Director Wray stated, “We’ve now reached the point where the FBI is opening a new China-related counterintelligence case about every 10 hours.” The threat is grave and our twentieth-century countermeasures, techniques and tradecraft are not appropriate for what many in the Intelligence Community deem the greatest threat to United States national security. Retooling, reimagining the intelligence recruitment cycle and modernizing the way that we approach the recruitment of sources is imperative.