DHS Seeks AI Biosurveillance Tools as Its Biodefense Office Splits

The Department of Homeland Security posted a request for information on May 18 asking industry to describe artificial intelligence (AI) tools capable of fusing biosurveillance signals from borders, supply chains, transportation networks, and agricultural systems into a single operational picture. Responses to the solicitation, catalogued as 70RSAT26RFI000019 by the DHS Office of the Chief Procurement Officer, are due June 5, and the agency says it will not answer questions before that deadline.

The urgency has a clear driver: the FIFA World Cup opens June 11 across North American host cities, the America250 centennial celebrations follow, and the 2028 Summer Olympics in Los Angeles adds the largest domestic biosurveillance burden in recent memory. Less visible is the structural context the solicitation enters. DHS is seeking an expanded AI biosurveillance capability at the precise moment its central biodefense coordination office is being dismantled.

The Capabilities DHS Is Polling For

Three capability clusters anchor the solicitation. Autonomous biological monitoring systems sit at the top: platforms that continuously scan environmental, epidemiological, and agricultural data without requiring manual analyst queries for each stream. Machine-learning (ML) recognition systems come second, built to identify unusual biological signatures against the background noise of routine public-health data. AI-powered anomaly analytics round out the core list, functioning as statistical engines that flag cross-source deviations and rank them by threat probability.

The data sources DHS wants those systems to ingest stretch well beyond the domain that biosurveillance programs have historically occupied. Border crossing indicators, supply-chain manifests, transportation flow data, and agricultural health reports all appear in the solicitation alongside more familiar inputs such as wastewater monitoring and hospitalization rates. A fourth capability category also appears: rapid confirmation methods for biological incidents, and biological incident risk analytics, suggesting DHS wants platforms that can model probable source, scale, and spread before any official declaration of a biological event.

The Science and Technology Directorate (S&T) overseeing the solicitation has run environmental detection programs for over a decade. According to the DHS Science and Technology Directorate biosurveillance program overview, S&T takes a system-level approach to integrating information into surveillance architectures, developing advanced detection systems, and maintaining a cross-domain focus on biological, chemical, and agricultural threats. What the new solicitation seeks is a machine-learned successor that fuses those legacy inputs with newer data categories those systems largely ignored.

A Splintered House, a Bigger Ambition

The CWMD Dissolution

The solicitation arrives inside a reorganization that biosecurity specialists consider poorly timed. The Trump administration’s FY 2026 budget proposes dissolving the Countering Weapons of Mass Destruction Office (CWMD), the unit that has led DHS’s biodefense coordination, chemical detection, and WMD response for more than a decade. The proposal would transfer 286 positions and $306.2 million in funding to other DHS components, per the FY 2026 CWMD Congressional Budget Justification published by DHS.

The splits are specific and spread wide. The National BioSurveillance Integration Center (NBIC), which consolidates biological threat data from partner agencies, moves to the DHS Office of Health Security. BioWatch and the Securing the Cities program shift to the Cybersecurity and Infrastructure Security Agency’s (CISA) Infrastructure Security Division. Detection equipment procurement fragments across Customs and Border Protection, the Coast Guard, the Transportation Security Administration (TSA), and the Secret Service. Each unit receives a piece of the former coordination structure; no unit holds the whole.

Biodefense Budgets in Retreat

The structural fragmentation is compounded by spending trends. According to the Council on Strategic Risks FY 2026 biodefense budget analysis, the Department of Health and Human Services (HHS) requested $19.44 billion for FY 2026, a $4.23 billion decrease from FY 2025 enacted spending. The Department of Defense (DoD) requested $4.02 billion, down $130 million from the prior year. The Department of Agriculture (USDA) requested $1.44 billion, a $54 million decline. Total government-wide biodefense requests for FY 2026 stood at $27.02 billion, against higher prior-year actuals across nearly every agency.

For the biosurveillance solicitation, the structural consequence is direct. S&T is asking vendors to describe systems spanning border data, supply-chain indicators, agricultural feeds, and transportation flows simultaneously, but the agency that would normally integrate those vendor responses across domains no longer has a single coordinating office for the work. The proposed buyer is a set of components that have never had to act as one unit on biosurveillance.

The Data Streams Behind the Ask

Biosurveillance programs built before 2020 relied primarily on syndromic surveillance (emergency department, or ED, visit patterns and pharmacy sales data) and environmental air sampling through the BioWatch sensor network deployed in high-risk urban areas. The new solicitation builds from those foundations but points upstream in the threat chain, toward data categories that older systems largely ignored. Supply-chain manifests carry cargo origins and handling conditions that can indicate deliberate contamination; border crossing indicators can precede outbreak signals by days; and transportation flow data maps where large populations are moving, toward which host cities, and how fast. The six data streams the solicitation seeks to fuse are:

• Wastewater surveillance – aggregate pathogen detection that typically precedes clinical case counts
• Hospitalization and ED reports – real-time clinical load by region and symptom cluster
• Border crossing indicators – traveler origins, volumes, and health-screening flags
• Supply-chain manifests – cargo provenance, cold-chain status, and biosafety flags
• Agricultural health reports – livestock and crop disease signals from USDA partner networks
• Transportation flow data – mobility patterns predicting population concentration in host cities

No single existing platform ties those six streams into one operational picture. That gap is the business case the solicitation is presenting to vendors, and the reason DHS needs responses by June 5 rather than sometime later in the year.

A Non-Governmental Operations Center Opens June 1

While DHS frames the solicitation as an investment for future events, one non-governmental coalition is already standing up biosurveillance capacity for the current one. Georgetown University and MedStar Health announced the launch of the Health Security Operations Center (HSOC) on June 1, ten days before the World Cup’s opening match. The center operates under the joint National Center for Health Security and Resilience, a Georgetown-MedStar academic health system partnership. More than 6.5 million soccer fans from over 100 countries are expected to travel to the United States, Mexico, and Canada for the June 11 to July 19 tournament.

The HSOC will draw on data from wastewater monitoring, hospitalizations, and real-time health reports gathered from host cities to deliver actionable information to health officials, per Georgetown University Medical Center’s HSOC launch announcement. More than 350 organizations and individuals, including hospital emergency managers, state and local health officials, and federal agencies, have already enrolled to receive daily situation reports. Publications begin June 4. The coalition supporting the HSOC spans more than 30 organizations across academia, public health, technology, and communications.

Mass-gathering events like the World Cup require the kind of coordinated, multidisciplinary situational awareness that no single institution or jurisdiction can provide alone.

Rebecca Katz, PhD, MPH, director of Georgetown’s Center for Global Health Science and Security and director of the HSOC, made that case in a statement released this month. The center is explicitly independent: it carries no official government program status and no FIFA sanction. That independence signals how large the gap between federal biosurveillance commitments and non-governmental health security capacity has grown in the current budget environment.

Farm to Threat: The Agriculture Layer

The inclusion of agricultural health reports among the solicitation’s target data sources reflects a DHS-USDA collaboration that has intensified sharply as AI expanded the options available to threat actors. Ashley Grant, senior health security and biodefense advisor at DHS’s office of health security, told lawmakers in February of this year that advances in biotechnology and AI have made it easier for state and non-state actors to develop and deploy biological agents capable of devastating crops, livestock, and food production, threatening economic stability and national security.

The congressional hearing at which Grant testified followed a September 2025 subcommittee session convened after Chinese nationals allegedly attempted to smuggle a potentially harmful fungus into the United States, according to the House Homeland Security Committee’s February agroterrorism hearing notice. The incident converted an abstract agroterrorism concern into a documented operational case study and drove deeper formal engagement between DHS and USDA on agricultural biosurveillance tools and protocols.

The two departments have shared research infrastructure for years. The National Bio and Agro-Defense Facility (NBAF) in Manhattan, Kansas, built by DHS’s S&T Directorate and transferred to USDA, provides biosafety-level-4 containment for large livestock research, the only such facility in the United States according to USDA’s program documentation. S&T also runs the Enhanced Passive Surveillance program, which integrates multiple livestock-health data streams in real time to identify disease outbreak trigger points and alert officials before a situation escalates.

The new solicitation extends that shared model toward a fully integrated alert architecture. Under the system DHS describes, a pathogen introduced through a cargo shipment and spreading through a regional livestock population should trigger flags simultaneously in supply-chain manifests, agricultural health data, and transportation corridor indicators. When those three streams converge on the same anomaly, that convergence is the signature the AI system is designed to catch before any single-stream detection would fire on its own.

The Dual-Use Problem at the Heart of the RFI

DHS is seeking AI tools to detect biological threats faster, but the agency has also been tracking a parallel concern for several years: the same AI capabilities that speed biosurveillance also make it easier to develop the threats those tools are designed to find. DHS’s CWMD office produced a report in April 2024, compiled under Executive Order 14110 on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, that assessed how AI intersects with chemical, biological, radiological, and nuclear threats across both offensive and defensive dimensions.

Biological Design Tools (BDTs) – AI systems that allow users to design novel organisms, characterize proteins, or model pathogen behavior – sit at the center of that concern. The CWMD report framed the governance challenge as finding the right balance between containing risk and fostering innovation, a balance that shifts as access to foundation models becomes cheaper and more widespread. Grant’s February congressional testimony updated the threat picture directly: AI has broadened the field of credible bioterrorism actors from primarily state programs to include well-resourced non-state groups, particularly for agricultural and food-supply targets.

The June 5 response deadline will tell DHS what industry believes is deployable now. If the post-CWMD command structure – Health Security, CISA, and S&T operating as separate principals – can integrate vendor responses into a coherent biosurveillance architecture, the tools have a buyer in time to stress-test them before the 2028 Los Angeles Olympics places the largest biosurveillance burden on the system it has ever faced. If the fragmentation baked into last year’s budget request proves harder to bridge than its architects assumed, the most capable AI biosurveillance platform the industry can offer will have no single agency responsible for fielding it.