# The Austin Surveillance Field ## The Instruments Were Already in the City In September 2012 the Austin City Council disposed of two adjacent agenda items. One sustained the **[[wiki/Austin Regional Intelligence Center|Austin Regional Intelligence Center]]**, the Central Texas fusion node organized through interlocal agreement beginning in 2010 among the city, surrounding jurisdictions, Travis and Williamson Counties, and UT Police, and equipped at its founding with criminal-justice information systems and **i2 link-analysis software**. The other authorized Austin's entry into the **[[wiki/Law Enforcement Analysis Portal|Law Enforcement Analysis Portal]]**, a North Texas repository and analytical environment operated through the [[wiki/North Central Texas Council of Governments|North Central Texas Council of Governments]] that aggregates Records Management System data across otherwise separate jurisdictions and supports link analysis, mapping, and cross-jurisdictional correlation — an architecture Dallas had already been building since 2008 through its own fusion initiative, **MOSAIC**. Two line items, transacted in a single meeting, and between them a person's Austin records became reachable from two hundred miles north. The local intelligence node was **federated by construction**, not by later drift. That is one instrument among many, and the density of the rest is what makes the city worth the study. Within the same metropolitan area sit **[[wiki/Applied Research Laboratories at UT Austin|Applied Research Laboratories at UT Austin]]**, a Navy University Affiliated Research Center conducting classified government research since 1946 and holding a Navy vehicle valued at roughly **$1.1 billion over ten years** covering sensors, signal processing, geolocation, cybersecurity, content understanding, and artificial-intelligence work explicitly directed at patterns that might predict terrorist or cyber attack; the Army UARC lineage of the **[[wiki/Institute for Advanced Technology|Institute for Advanced Technology]]**, whose chemical-biological countermeasures program reported to the UT System in the 2003–2005 accountability cycle that it would develop biological-agent sensors, build an archival Texas disease dataset, and **"conduct surveillance in real time of patients entering emergency medical facilities"**; a **[[wiki/Defense Threat Reduction Agency|Defense Threat Reduction Agency]]** biosurveillance award to UT researchers that computed a six-dimensional trust score for every incoming information source; a documented research ancestry in which **[[wiki/Google Flu Trends|Google Flu Trends]]** functioned as a virtual epidemiological surveillance provider, with Austin among eight Texas cities carrying city-level search-derived influenza signal by 2010; **[[wiki/HID Global|HID Global]]**, headquartered on Center Ridge Drive, as one of four publicly identified prime performers on an ODNI biometric-authenticity program; the **[[wiki/PECOS|Center for Predictive Engineering and Computational Sciences]]**, founded on a $17 million NNSA award whose entire disciplinary premise is reconstructing hidden states from incomplete observations and quantifying how far the reconstruction may be trusted before anyone acts on it; and **[[wiki/Texas Advanced Computing Center|TACC]]**, whose Horizon system now supplies exaflop-class computation to run those reconstructions at population scale. None of that is a program. It is an accumulation. And its consequence is that by the middle of the 2010s a person did not have to be identified, cultivated, transported, or enrolled to become useful to it. A person could simply **arrive** — fly commercially, carry a phone, run a search, drive through a traffic stop, walk into an emergency department, generate a laboratory result, live inside a wastewater catchment, post publicly, pass a building — and become observable along seven or eight independent axes without ever encountering a recruiter, signing a consent form, or perceiving a single unified apparatus. Federal insiders have already established that the selection mechanism operates without criminal predicate: **[[wiki/Quiet Skies|Quiet Skies]]** placed roughly five thousand Americans who were neither suspects nor watchlisted under covert physical surveillance by armed federal teams. A January 2026 Government Accountability Office report establishes that watchlist status follows a person into an ordinary traffic stop and out again into a fusion-center notification. Austin's own 2018 warrant record establishes that investigators here sought unknown identities by querying the spatial and behavioral exhaust of entire user populations rather than by naming a suspect first. The operative question in such an environment is not who was targeted. It is **what is addressable**. Because that is a claim about architecture rather than about intent, it has to be argued at explicit evidentiary tiers, and this piece uses them consistently. **Established** means documented in primary institutional, judicial, congressional, or audit records. **Strongly indicated** means multiple independent documents converge on the same structure without a single dispositive source. **Plausible** means the mechanism exists and the motive exists but the specific instantiation is undocumented. **Unresolved** means the question is genuinely open and the missing evidence is identifiable. **Unsupported** means no public evidence presently exists. **Contradicted** means the record affirmatively cuts against the claim. Plausibility derived from motive or infrastructure is never silently promoted into fact, and the absence of a document is never treated as proof that the underlying activity did not occur — particularly where classified reporting channels exist that the open record cannot see. The discipline runs in both directions. ## The Whistleblowers Appear Upstream, at the Transportation Layer The most probative disclosures do not come from a laboratory, from DTRA's Surety BioEvent project, or from any Austin institution. They appear in the federal aviation and traveler-screening infrastructure that attaches surveillance to a person **before that person reaches any city at all**. In 2014 the U.S. Office of Special Counsel received a whistleblower disclosure concerning **TSA's [[wiki/Secure Flight|Secure Flight]]** program, alleging that a risk-based Secure Flight rule created a security vulnerability. OSC referred the matter to the Department of Homeland Security, whose Inspector General interviewed the discloser and senior TSA personnel and examined the underlying rules and program documentation; portions of the resulting review remained withheld as Sensitive Security Information. That disclosure concerned efficacy rather than impropriety, but it establishes something structurally important and **established**: Secure Flight was already executing algorithmic risk determinations about individual passengers under rules whose operation was opaque and consequential enough that insiders went outside the ordinary chain of command to challenge them. Four years later the disclosure trail became directly dispositive. In July 2018, federal air marshals and internal TSA documents exposed **Quiet Skies**, a previously undisclosed domestic passenger-surveillance program in which armed undercover teams shadowed ordinary U.S. citizens who were **not criminal suspects and were not in the Terrorist Screening Database**, documenting behavior minute by minute — conversation partners, device use, apparent alertness, lavatory visits, post-arrival movement, and other mundane physiological and behavioral observations. Travelers were not notified. Selection required no individualized criminal predicate: all citizens returning from abroad were screened against rules incorporating **travel history, travel patterns, affiliations, watchlist information, and other government databases**, and one documented criterion concerned reservation information — an email address or telephone number — associated with a watchlisted person. Selected travelers could remain under coverage for as long as ninety days or three encounters. In August 2018, TSA briefers told congressional staff that roughly **5,000 Americans** had already been monitored without producing a single case warranting further scrutiny; in September, Administrator David Pekoske acknowledged in testimony that the program had foiled no threat and produced no arrest, and that the resulting behavioral records could be retained for two years. By December 2018, TSA had curtailed the most intrusive collection. The trail did not end there. On **September 30, 2025**, the Senate Homeland Security and Governmental Affairs Committee convened a full-committee hearing titled _Examining the Weaponization of the Quiet Skies Program_, taking testimony from **Senior Federal Air Marshal Mark Crowder** — a twenty-year air marshal with prior service in the Marine Corps, as a uniformed CIA police officer, and as a Florida deputy sheriff — alongside Empower Oversight, Racket News, the American Enterprise Institute, and the American-Arab Anti-Discrimination Committee. Crowder testified that his wife had been flagged under **"Special Mission Coverage"** because she shared an itinerary with someone the agency had mistakenly identified as having entered the Capitol on January 6; TSA subsequently acknowledged the mistaken identity. Committee materials placed additional monitored individuals into the record, and the program was terminated by the Secretary of Homeland Security before the hearing. The partisan framing of that hearing requires its own evidentiary evaluation and does not carry the argument. What carries the argument is narrow, **established**, and sufficient: federal insiders disclosed a system in which people who were neither criminal suspects nor watchlisted could be **algorithmically selected and physically shadowed without their knowledge**, and the selection criterion could be nothing more than an identifier shared with someone else. That is the entire mechanism the strong hypothesis requires, and it operates without any Austin-specific recruitment. A person does not need to be a designated suspect. A person does not need to know anyone suspicious in the ordinary social sense. An **identifier, an association, an itinerary shape, or a behavioral rule** suffices to elevate the individual into a different computational state, after which the human becomes the object of persistent collection. ## The Watchlist Does Not Stay at the Airport The screening state migrates into ordinary local encounters, and the mechanism is now documented at the level of federal audit. On **January 12, 2026**, the Government Accountability Office publicly released _Terrorist Watchlist: FBI Should Improve Outreach Efforts to Nonfederal Users_ (GAO-26-108650), the public version of a sensitive report issued in August 2025. GAO found that nonfederal officers routinely query encountered individuals against watchlist information during ordinary interactions — **traffic stops** are the canonical example — through state database systems that return an alert on a potential match, after which the officer contacts the FBI's **Threat Screening Center** for resolution. Nonfederal entities in all fifty states, the District of Columbia, and four tribal governments reported encounters between fiscal years 2019 and 2024, though GAO also found that such entities account for only about five percent of watchlist encounters overall and that reporting is inconsistent because most interviewed agencies had received no training on watchlist policy. When an encounter resolves to a confirmed positive match, the architecture expands rather than terminates. The Threat Screening Center can notify the relevant fusion center through DHS's **[[wiki/Homeland Security Information Network|Homeland Security Information Network]]**, and the notification can carry **date and time, encounter location, identifying information, watchlist handling code, reason for the encounter, encountering agency, vehicle information, and additional biographical detail**. Officials at seven of the nine fusion centers GAO interviewed said they would then query additional databases about the watchlisted person, frequently sharing results with DHS, the FBI, and state organizations; some aggregated multiple encounters into broader analytical products describing watchlisted activity across their geographic area. The sequence — **travel identity → passenger screening → watchlist state → local encounter → fusion-center notification → additional database queries → regional intelligence product** — is documented end to end, and no recruiter appears anywhere in it. The subject merely moves. ## Loder: An Insider Inside the Fusion-Center and Task-Force Ecology The strongest insider match to the fusion-center layer specifically is **Maine State Trooper George Loder**, whose case matters precisely because he sat at the seam between a DHS-recognized post-9/11 fusion center and an FBI Joint Terrorism Task Force. Assigned to the **Maine Information and Analysis Center** in 2013 and posted almost immediately to the JTTF in Portland, Loder alleged that MIAC unlawfully retained personally identifiable information, collected on peace and environmental activists and on people at a camp for Israeli and Arab teenagers, maintained information about gun purchasers, and pressured him to disclose restricted task-force information. In December 2022, a nine-person federal jury in Portland found that the Maine State Police had violated the state whistleblower statute in retaliating against him and awarded **$300,000** in compensatory damages. The evidentiary discipline here is essential and is routinely botched in secondary commentary. The verdict **established retaliation**; it did not adjudicate every underlying surveillance allegation, which remain at the **strongly indicated** tier — corroborated by the state's own subsequent commitment to an external review of intelligence-gathering practices and by a legislative defunding attempt, but not judicially proven item by item. What Loder supplies to this argument is the demonstration that an intelligence officer operating inside the federally supported counterterrorism information-sharing architecture alleged, under oath and at professional cost, that such a node had crossed from threat intelligence into the retention of information about lawful people. That is a documented failure mode of the architecture rather than a hypothesis about it. ## Austin Already Possessed the Receiving Node, and It Was Federated Austin's corresponding institution is the **Austin Regional Intelligence Center**, organized through interlocal agreements beginning in 2010 among Austin, surrounding jurisdictions, Travis and Williamson Counties, UT Police, and other participating agencies. Procurement records show that ARIC's creation was technological as much as organizational, involving criminal-justice information systems and **i2 link-analysis software** for relationship analysis and data integration. Its function in this architecture is precise: it is the local node through which national screening and intelligence information becomes **operationally meaningful about a person physically present in Central Texas**. The national fusion model had already generated severe oversight findings before ARIC matured. A bipartisan two-year review by the Senate Permanent Subcommittee on Investigations, released in 2012, concluded that federally supported fusion centers frequently produced intelligence that was useless, irrelevant, unrelated to terrorism, or potentially inconsistent with civil-liberties and Privacy Act protections; of 610 intelligence reports examined, **188 were never published**, often because they contained no useful intelligence or created privacy problems, and investigators found instances in which problematic U.S.-person information nonetheless remained stored inside DHS systems. Austin then produced its own unusually granular demonstration. The 2020 **BlueLeaks** breach exposed more than ten gigabytes of records associated with ARIC. Provenance discipline matters: BlueLeaks was the product of a hack, not a protected disclosure, and it should never be characterized as whistleblowing. What the material established is that ARIC operated a substantial **[[wiki/Threat Liaison Officers|Threat Liaison Officer]] network** extending beyond sworn law enforcement into other institutions and private-sector participants, that liaison officers could submit suspicious-activity observations into the fusion environment, and that the categories of reportable behavior were broad enough to encompass observation, surveillance-like conduct, and unusual social-media activity. Subsequent examination found records concerning people whose behavior admitted obvious innocent explanation — customers attempting to mail a package to Lebanon became the subject of a suspicious-activity report containing identifying information and video, and an ARIC analyst eventually determined there was **no criminal or terrorism nexus**, yet the records persisted in local intelligence files years afterward; men characterized as Middle Eastern whose parking-lot conduct appeared to include prayer generated a similar analytic finding of no violation and a similar circulation "for situational awareness." Further reporting found lawful activist activity categorized in intelligence products, along with monitoring of demonstrations and of an online Juneteenth event during the summer of 2020. None of that generalizes into a claim that Austin residents were broadly surveilled by ARIC, and it should not be stretched that far. It establishes something narrower and more useful: **the threshold between ordinary life and intelligence collection was demonstrably permeable, and permeable in the direction of persistence** — an analytic finding of no nexus did not reliably purge the record. That permeability is worth measuring against the center's own governing instrument, which is unusually elaborate on paper. The ARIC interlocal agreement binds collection to 28 CFR Part 23 and a reasonable-suspicion standard, forbids seeking or retaining information about a person or organization **solely on the basis of religious, political, or social views, participation in an organization or event, race, ethnicity, citizenship, place of origin, age, disability, gender, or sexual orientation**, and expressly provides that those factors shall not be treated as creating suspicion. It establishes a trained Privacy Officer charged with reviewing every analytical product before dissemination, a Privacy Policy Advisory Committee seated with a community advocate and an information-privacy advocate, and an external audit function subject to review by the Austin Public Safety Commission, which reports to City Council. The BlueLeaks material is therefore not evidence that Austin lacked privacy governance. It is evidence of the distance between a governance document and a retention practice — which is the more instructive finding, because the document was among the better ones. The structural fact more consequential than any of this is that the node was never merely local. ## LEAP: The Bridge That Outlived Its Jurisdiction The **Law Enforcement Analysis Portal** deserves separate treatment because it is the component of the Austin architecture that has traveled furthest from the conditions under which the city agreed to join it, and because its trajectory demonstrates in miniature what happens to public information-sharing infrastructure over a twenty-year horizon. LEAP began as a project of the **North Central Texas Council of Governments' Criminal Justice Program**. A January 2010 partnership announcement describes the project as having launched roughly three years earlier — placing its origin near 2007 — and as already supporting **more than 200 law enforcement agencies in more than 75 counties across multiple states**. It was designed to aggregate Records Management System and related criminal-justice data from jurisdictions whose systems could not otherwise communicate, and to supply crime analysis, link analysis, mapping, and cross-jurisdictional correlation on top of that aggregation. Crucially, it was federally interfaced almost from the start. A Department of Homeland Security **[[wiki/DHS Law Enforcement Information Sharing Service|Law Enforcement Information Sharing Service]]** briefing lists **"LEAP (NCTCOG)"** among its established sharing partners alongside San Diego's [[wiki/ARJIS|ARJIS]], [[wiki/LInX|LInX Northwest]], [[wiki/OneDOJ|OneDOJ]], the Los Angeles Sheriff's Department, [[wiki/Texas Data Exchange|TDEx]], and National Capital Region LInX, with agreements for **[[wiki/National Data Exchange|N-DEx]] and [[wiki/Nlets|NLETS]]** anticipated in fiscal year 2010; LEAP in turn advertised to its own users access to national holdings including **ONEDOJ and [[wiki/ICEPIC|ICEPIC]]**. It also ingested commercial patrol-level streams, entering a partnership in January 2010 with [[wiki/COPsync|COPsync]] for **citations, offense and incident reports, and GPS coordinates associated with law enforcement activity**, collected in real time at the point of contact. The North Texas side of that architecture had its own fusion node. Dallas developed **MOSAIC**, the Metro Operations Support and Analytical Intelligence Center, which appears in municipal records by 2008 as an intelligence and analytical fusion center and lists **LEAP among the activities under development in its first year**, alongside analytical tooling and federal, state, local, and public-private relationships. That establishes a real institutional relationship between the Dallas fusion environment and the NCTCOG portal without establishing that the two were the same system. IARPA's later and entirely unrelated use of the MOSAIC acronym — **Multimodal Objective Sensing to Assess Individuals with Context**, launched September 2017 — is a genuine terminological recurrence with **no demonstrated organizational lineage** to the Dallas program, and it should be preserved at that tier rather than promoted into a connection. That is the object Austin joined in **September 2012**, on a City Council agenda where the adjacent item concerned ARIC sustainment — the second interlocal instrument, entered by the partner agencies across the fall and winter of 2012 into 2013, that funded the technology and systems necessary for continued operation of the fusion center. The two actions are properly read as one architectural decision. Austin was simultaneously capitalizing the technology layer of its own regional intelligence node and attaching that node's data environment to a multi-state aggregation already interfaced with federal exchange systems. The agreement was not a mutual-aid courtesy with Dallas. It was **an entry point into national criminal-justice information sharing routed through a council of governments**, which is a form of federal integration that requires no federal agreement with the City of Austin at all. The 2015–2018 window contains no discrete Austin LEAP action that surfaces in the open record, and that absence should be stated plainly rather than filled: municipal procurement records, NCTCOG contract files, and APD's own systems inventory would resolve whether the city's participation was renewed, expanded, or allowed to lapse in those years, and that is a **tractable, identifiable gap** rather than a mystery. What the window does contain is the governance change that matters most. In **2017**, the ARIC partner agencies amended both the original agreement and the sustainment funding agreement to authorize the City of Austin to execute **[[wiki/Partner Equivalent Agency Agreements|Partner Equivalent Agency Agreements]]** — collateral instruments allowing **non-governmental agencies** to assume the obligations and receive the benefits of full partner agencies, conditioned on the entity operating a full-service law enforcement component staffed by Texas-licensed peace officers, and requiring approval by the Executive Board, its chair, and City Council. The private institutions that entered through that mechanism appear on the later restated agreement: **St. Edward's University and Concordia University**, alongside a partner roster that by then had grown from ten founding agencies to twenty-six, including four independent school district police departments, four university police departments, and a community college. The fusion center did not merely persist across the 2015–2018 period. It **widened its aperture and opened a formal channel for non-governmental membership**. LEAP itself did not sunset either — it was privatized and then nationalized, and the name survived the transition. **[[wiki/Forensic Logic|Forensic Logic]]**, founded in 2003 in Walnut Creek, California, built a law-enforcement search engine it also called the **LEAP Network**, capable of indexing structured and unstructured police data under natural-language query and marketed by its chief executive as bringing "Google to law enforcement." Oakland became its first major customer in 2012, ingesting ShotSpotter acoustic detections, NIBIN ballistic linkages, and automated gun-trace submissions, and querying the portal roughly **70,000 times a month** by 2017. That same year Forensic Logic acquired **[[wiki/COPLINK|COPLINK Detect]] from IBM**, the 1998-vintage Knowledge Computing product that had achieved nationwide adoption between 2006 and 2015, and merged it with its own search technology as **[[wiki/COPLINK|COPLINK X]]** in 2021. On **January 3, 2022**, ShotSpotter acquired Forensic Logic for approximately **$36.2 million** in cash, stock, and contingent consideration. The company rebranded as **[[wiki/SoundThinking|SoundThinking]]** in April 2023 and renamed the product **[[wiki/CrimeTracer|CrimeTracer]]**, which now advertises **more than a billion records across thirty-plus data sources** — records management, computer-aided dispatch, license plate readers, ShotSpotter detections, jail management, court and warrant records — sold as statewide subscriptions including Tennessee in 2021 and a Massachusetts contract exceeding $4 million across forty-four months in December 2022. In August 2023 the same company began absorbing the engineering team of **Geolitica**, formerly PredPol. Whether NCTCOG's portal was built on Forensic Logic's platform or was a parallel governmental system sharing a name and a function is **strongly indicated but not established** on the public record — the name is identical, the function is identical, the periods overlap, and Forensic Logic dates its own LEAP Network to 2003, but no document reviewed here states the relationship directly. NCTCOG procurement and contract files would settle it, and the two readings carry different implications: under one, Austin's 2012 vote attached the city to a commercial corpus that later became CrimeTracer; under the other, it attached the city to a governmental predecessor whose regional function was subsequently displaced by that corpus. Either reading produces the same structural conclusion, which is why the ambiguity does not paralyze the analysis. That conclusion is the one worth carrying forward. The regional public information-sharing layer did not terminate under the oversight pressure that fusion-center criticism generally anticipates. It **consolidated upward into private ownership**, where retention schedules, participating-agency rosters, access rules, and audit trails are governed by subscription agreements rather than by council votes, and where records aggregation now ships bundled with acoustic gunshot sensing, plate reading, and the absorbed remains of predictive policing inside a single vendor stack. A record that Austin contributed to a council-of-governments node in 2012 plausibly sits today inside a billion-record commercial index sold to jurisdictions that never negotiated with Austin and answer to no one in Travis County. ARIC persists under an interlocal agreement with a Privacy Officer and a City Council audit path; the data environment it participates in has migrated somewhere those instruments do not reach. This is precisely the transition **[[wiki/IC ITE|IC ITE]]** executed inside the classified enterprise — from human-negotiated bilateral sharing toward a machine-queryable common corpus — performed here in the commercial market, at municipal scale, with no equivalent oversight literature attached to it and no single moment at which any public body was asked to approve the result. ## Austin Built the Biological Layer a Decade Before BioEvent The intelligence layer is only one half of the field, and the biological half is chronologically older than the Austin literature usually admits. The **[[wiki/National Biological and Chemical Countermeasures Program|National Biological and Chemical Countermeasures Program]]**, created in 2000 within the Army UARC environment associated with UT's **Institute for Advanced Technology**, joined UT System campuses with the Texas Department of Health, the Texas National Guard's **6th Civil Support Team**, emergency-management organizations, the Institute for Defense Analyses, and other defense participants. A 2002 report on the Army's UARC chemical/biological countermeasures work describes its three principal elements as **sensor development, medical countermeasures, and communications**, and a surviving UARC program briefing enumerates its components as _Sensors and Situation Awareness_, _Medical Countermeasures_, _Archival Data Set and Disease Surveillance_, _Biosurveillance_, _Intelligent Software Agents_, and an **Interface with INTEL-FBI**, with Central Texas FBI personnel among the consortium and a proposed technology demonstration inside a 72,000-square-foot underground facility at Fort Hood. The UT System's own accountability reporting from the 2003–2005 period is more explicit than any external interpretation could be. Under **Countermeasures to Biological and Chemical Threats**, UT Austin reported programs to develop biological-agent sensors, develop vaccines, establish an archival dataset of disease in Texas, and **"conduct surveillance in real time of patients entering emergency medical facilities."** A 2002 technical description of the program explains the mechanism: large-scale, real-time symptomatic diagnoses from patients entering emergency medical facilities were to be collected electronically and transmitted to an archival facility for identification of emerging disease, with the explicit objective of compressing outbreak-recognition time from a week or more toward roughly twenty-four hours. This is **established**, self-described, and decisive for chronology. Austin's biosurveillance genealogy does not begin with DTRA's BioEvent award in 2013 or 2014; the architecture — **patient encounter → electronically captured health observation → centralized archive → anomaly recognition → identification → situational awareness → emergency response** — was operating as a funded institutional objective more than a decade earlier. The same environment investigated the built environment itself as a sensing medium. A surviving UARC research presentation describes surface sampling in **thirty dormitory rooms, three surfaces per room**, followed by broth culture, amplification with multiple pathogenicity-island primer sets, and multiplex probe analysis, with the finding that the targeted sequences were not ubiquitously distributed in those spaces; the accompanying technical work was funded under **U.S. Army Research Laboratory contract DAAD17-01-D-0001**, pursued through the Institute for Advanced Technology with Army chemical-biological personnel and a Georgetown, Texas biotechnology firm, and aimed at a fieldable multiplexed detector alerting military, health, and government users. The envisioned deployment environments were **military operations, hospitals, schools, offices, food settings, and general civilian buildings**. What the record does not disclose is which dormitory supplied those rooms or what consent protocol governed the sampling, and neither fact can be supplied by inference. The correct tier is therefore precise: environmental dormitory-room biosampling by the UT UARC program is **established**; covert sampling of identified students is **unsupported**. But the architectural implication survives the uncertainty intact, and it is the more consequential point. The object of observation had already begun shifting away from the enrolled research participant toward **the environment through which unenrolled populations leave biological traces**. ## The Legal Hinge: Observation Is Not Necessarily Research This is where the entire question reorganizes, and it is the single most important distinction in the article. UT's research-policy materials reproduce the federal Common Rule provisions under which several categories of deliberate governmental observation are **not research** for regulatory purposes: authorized public-health surveillance necessary to identify, monitor, or investigate disease signals; criminal-justice collection authorized by law or court order; and **authorized operational activities in support of intelligence, homeland-security, defense, or other national-security missions**. That exclusion is not a license to conduct undisclosed experiments on people, and it does not suspend constitutional limits or other privacy statutes. What it establishes is far more structurally significant: **a person can be intentionally, systematically, and lawfully observed by a government system without ever becoming a "research subject."** The contrast with actual human-subject research under the same sponsor is stark. DTRA's 2014 _Fundamental Research to Counter Weapons of Mass Destruction_ solicitation required proposed human-subject work to identify the source of subjects, document local Institutional Review Board review, comply with DoD and federal protections, provide consent materials, and obtain separate approval from DTRA's **Research Oversight Board** before any human-subject research could begin. UT's contemporary policy states that it does not conduct classified human-subject research and does not conduct human-subject research involving chemical or biological agents as defined by the cited DoD directives. Consequently the question "did BioEvent experiment on unwitting Austin residents" is the wrong question, and answering it in the negative — as the public record does — resolves almost nothing. BioEvent never needed to experiment on a person in order to make **human-generated information part of its inference system**. The regulatory boundary and the observational boundary are simply not the same boundary, and the gap between them is where the surveillance field lives. ## Barber's Genealogy and the Migration of the Trust Problem The public grant record for **[[wiki/Suzanne Barber|Kathleen Suzanne Barber]]** ties together domains that appear unrelated when examined one project at a time. Her history runs from the original biological/chemical countermeasures consortium and _Information Systems for Distributed Chem-Bio Incident Response_ through DARPA distributed intelligent-agent research; from 2005 through 2008 she led the approximately **$900,000 Office of Naval Research project "Support for Predicting Improvised Explosive Device Attacks"**; from 2006 through 2008, roughly **$627,736** for _Rational Explanation of Beliefs, Intentions, and Threats_, connected to maritime-domain awareness; in 2009, a DHS award for the precisely titled **"[[wiki/Sensor and Social Networks Armed to Detect and Defend against Terrorist Attacks|Sensor and Social Networks Armed to Detect and Defend against Terrorist Attacks]]"**; IBM-funded _Identity Analytics and Visualization_ in 2012; an Army Operational Test Command project in 2013–2014; DTRA's **[[wiki/Surety BioEvent App|Surety BioEvent App]]**, formally performed from May 1, 2014 through April 30, 2017 at approximately **$2.77 million**, with the UT announcement appearing in September 2013 and accounting for the apparent one-year discrepancy; and subsequent work on identity-threat assessment and DHS **Office of Biometric Identity Management** research. The lineage — **distributed agents → chem-bio incident response → IED attack prediction → beliefs and intentions → sensor and social-network terrorism detection → biosurveillance source trust → biometric identity** — is not a career drifting between unrelated fields. It is a single methodological problem instantiated in successive substrates: heterogeneous observations must be assessed for reliability, fused into a hidden-state estimate, interpreted for threat significance, and used to select an intervention. BioEvent makes the epistemic core of that problem explicit in a way little else in the corpus does. The system computed a six-dimensional **trust representation for each information source — Identity, Expertise, Reputation, Experience, Authority, and Separation** — and applied goal-oriented multi-objective optimization to match source trustworthiness against a specific surveillance objective under constraints. Its predecessor influenza work had already established that a source performing well in one epidemiological regime could actively mislead the system in another. The operative question was therefore never _collect everything_. It was **who produced this observation, how reliable has that source been, does it carry independent information or merely echo another source, how much authority should it receive, and how well does it predict the biological state being estimated**. That is verification, validation, and uncertainty quantification translated out of mathematical physics and into heterogeneous human and open-source data — the same problem PECOS was founded to solve for systems that cannot be experimentally tested at full scale. Barber's professional activity during 2016–2017 shows how tightly this research abutted the operational architecture: her CV records presentations on **"Identity Intelligence" to CIA Digital Futures in Austin, "Belief in Identity" to the FBI in Austin, "Trust in Social Media" to DTRA, "Finding Trusted Sources for BioSurveillance" to DTRA, "Consumer Identity Management and Biometrics" to DHS's Office of Biometric Identity Management, and trusted-identity research at an NSA Innovation Forum**. That is an unusually concentrated intersection of biosurveillance, social information, identity, and national-security organizations in a single year. It is also, on the public record, **entirely lawful academic engagement**, and no public whistleblower, Inspector General finding, litigation, IRB disclosure, or grant record names Barber, Lauren Ancel Meyers, Andy Ellington, or the BioEvent team in connection with covert observation or unauthorized targeting. The significance is architectural and lies in the interfaces, not in an accusation the evidence does not support. ## Google Enters Twice, and Only One Entry Is Documented The consequential Google connection is not a building. It is **behavioral data**. Meyers and collaborators published work in 2012 on optimizing influenza-surveillance networks using historical disease surveillance together with internet-search information, in which **Google Flu Trends functioned as a virtual surveillance provider** evaluated mathematically against conventional physician-based reporting to determine which combination of sources and locations produced the most useful state estimate. By August 2010 the Texas influenza-surveillance handbook already listed **Austin among eight Texas cities** with city-level Google Flu Trends information available. The Surety BioEvent announcement explicitly invokes this lineage and proposed extending the trust framework across traditional reporting, open-source material, and social media. This is the literal instance of unwitting contribution, and it is **established**. People typed searches because they wanted information. They enrolled in nothing, met no recruiter, signed no consent form, and had no need to know that Meyers or Barber existed. At aggregate scale their search behavior became an epidemiological observable from which machines attempted to reconstruct the hidden biological state of a population, and a DTRA-funded project subsequently proposed to weight that observable by computed trust. The second Google entry is geographic and must be held at a lower tier. Google announced Austin as its next Google Fiber city on **April 9, 2013**, months before UT publicly announced BioEvent, and in **December 2014**, inside BioEvent's formal performance period, opened its downtown **Fiber Space in the former Austin Children's Museum** on Colorado Street in the Second Street district, describing it as a community gathering and demonstration venue rather than merely a sales location. Google's subsequent Austin office expansion moved a few blocks west along Second Street, further concentrating corporate presence in the same compact corridor immediately north of Lady Bird Lake. There is **no documentary evidence connecting the Fiber Space, the residential buildings in that corridor, or the physical fiber network to DTRA's Surety BioEvent program**, and that boundary should remain explicit and permanent. The overlap is worth recording only because it establishes the setting accurately: during BioEvent's operational period, a person living or moving through that corridor was simultaneously inside a city deploying Google Fiber infrastructure while UT researchers used **Google-derived behavioral information** as a biosurveillance input. Two Google stories intersect in geography and chronology; exactly one is documented as part of the research ancestry. ## Austin Then Produced the Literal Reverse-Search Dragnet Within a few years Austin became one of the country's more important demonstrations of **inferring an unknown person from data rather than querying data about a known person**. During the March 2018 serial-bombing investigation, investigators obtained a **Google geofence warrant** against SensorVault location information, seeking account information associated with devices present in defined geographic areas during relevant periods; subsequent legal analysis indicates the location data produced no useful lead toward the bomber. The technique nonetheless inverts the traditional investigative direction. Rather than beginning with a named suspect and asking where that person traveled, a geofence query begins with **space and time** and asks who was present, so that every device satisfying the geographic criterion — including devices belonging to people against whom no suspicion existed — initially enters the investigative field. The same investigation went further. Five **keyword warrants** were issued by the U.S. District Court for the Western District of Texas, Austin Division, on **March 14, 2018**, directed at Google, Microsoft's Bing, and Yahoo, seeking IP addresses and account information for users who had searched the targeted residential addresses and terms associated with bomb construction, including more than half a dozen variations on _pipe bomb_ and _PVC bomb_ over roughly a two-month window; one request also sought users who had searched a victim's address using **Google Maps or Waze**. What the tech companies returned remains under seal. These warrants postdate BioEvent's conclusion and there is **no evidence tying them to DTRA research**. Their significance is that they establish, in Austin specifically, an operational principle: **a person can become visible to law enforcement not because the government knows who they are, but because their digital trace satisfies a place, a time, or a query criterion**. The target is inferred from the data rather than specified before collection. ## Biological Space Yields to the Same Inversion The biological equivalent works identically, which is why the two halves of this field are not merely adjacent but structurally identical. Traditional clinical research begins with known subjects; surveillance begins with an **anomalous signal** and works backward toward its source. Contemporary Texas **syndromic surveillance** institutionalizes the model the UARC program described in 2002. A patient arrives at an emergency department, information is entered into the electronic health record, selected elements are transmitted to the statewide system, algorithms search the stream for abnormalities, and alerts are generated for public-health investigation, with no separate data-entry step performed by the patient and no change to the normal clinical encounter. Texas DSHS states that participating facilities transmit **individual-level data**, that the statewide ESSENCE platform holds emergency-department information alongside poison-control data, death records, and environmental feeds, and that appropriately authorized health departments can work with providers to **re-identify patients when an alert produces a case investigation**. The record is protected health information throughout; it is also, functionally, an early-warning sensor reading. **Wastewater surveillance** extends the inversion into a domain with no encounter at all. Beginning in 2020, UT Austin investigators sampled Austin's two largest treatment plants to estimate SARS-CoV-2 prevalence, with the explicit advantage that sewage captures a population vastly larger than the subset who voluntarily seek testing. The resolution then tightened: UT and Austin Water collected from **five manholes on and west of the UT campus**, coupled results with TACC analysis and city sewer maps, and planned to determine which buildings drained into which pipes. A **$150,000** Texas Division of Emergency Management grant resumed funding in 2022 for sampling at the Walnut Creek and South Austin Regional plants and proposed inexpensive samplers at multiple UT locations to determine whether **specific buildings or residence halls** showed rising viral signal, with researchers noting that wastewater could eventually be interrogated for additional pathogens and human disease biomarkers. Austin Public Health separately used Biobot monitoring for Travis County variant surveillance. The sewer user is not clinically enrolled, and the measurement is not an individual diagnostic: the catchment is the observational unit and the initial output is a **collective biological state**. Finer sampling progressively reduces geographic uncertainty — metropolitan area, to neighborhood, to campus, to building — and the Austin documents do not describe resolving signals back to named individuals. But the primitive is unmistakable and identical to the geofence primitive: **unknown people → involuntary emission → aggregate sensor → anomaly detection → spatial localization → targeted follow-up**. The early dormitory pathogenicity-island work sits at the same coordinate from the opposite direction, treating occupied built environments as collection surfaces. ## Identity Becomes a Computational Primitive, in Austin The overlay most often missed in accounts of this architecture is that the **Identify** stage was being industrialized in Austin at the same moment. **[[wiki/IARPA Odin|IARPA Odin]]** — a program of the Office of the Director of National Intelligence, initiated through a March 2016 Proposers' Day and a June 2016 solicitation and publicly announced in October 2017 — developed **biometric presentation-attack detection** for face, iris, and fingerprint systems using deep learning, computer vision, visible and multispectral imagery, hardware and software sensing, normalcy modeling, and anomaly detection, with a final performance target of approximately **97 percent average attack-detection accuracy at a 0.2 percent false-alarm rate**. That specification is itself a formal statement about the asymmetric cost of mistaken inference. The program was internally divided between **Thor**, an unclassified defensive performer environment, and **Loki**, a classified performer base tasked with discovering vulnerabilities deliberately withheld from Thor participants — an elegant adversarial architecture in which the defensive research community is not permitted full knowledge of the offensive research community. Among only four publicly identified prime performers was **HID Global, headquartered at 611 Center Ridge Drive in Austin**, whose commercial domain is precisely credentials, authentication, biometrics, and access control. IARPA's separately named **MOSAIC — Multimodal Objective Sensing to Assess Individuals with Context**, publicly launched September 11, 2017 — pushed the sensing target further inward, pursuing unobtrusive, passive, continuous sensing of psychological, cognitive, physiological, behavioral, and contextual characteristics through mobile devices, wearables, software, and environmental sensors, ultimately assembling data from more than 1,400 volunteer participants. The documented scope was an **Intelligence Community workforce research program**, not a general-population surveillance effort, and that boundary is real and must be maintained. Its architectural implication survives the boundary: the question being asked was whether persistent heterogeneous sensor streams could be fused into **machine assessments of changing human state**. Austin holds the consented civilian analogue of exactly that architecture in UT's **[[wiki/Whole Communities–Whole Health|Whole Communities–Whole Health]]** program, which received IRB approval for a five-year longitudinal cohort and by the end of 2024 reported 497 participants from 160 families equipped with wrist-worn devices, smartphones, and environmental sensing beacons placed inside their homes, generating sleep, activity, stress, travel-behavior, and environmental measurements. That program is explicitly consent-based and community-engaged; it belongs in this article not as an object of suspicion but as the demonstration that **longitudinal human state estimation from continuous sensor streams is a mature, ordinary, locally available capability** whose difference from covert collection is legal and procedural rather than technical. The gradient reaches its present terminus in the semantic interior. In 2023, UT Austin researchers demonstrated a **noninvasive semantic decoder** capable of translating fMRI activity, recorded while participants listened to or imagined stories, into continuous language representations using transformer-based methods. The correct tier here is unusually important because this finding is routinely inflated: the system requires **cooperative participants and extensive individualized training data**, it does not produce verbatim access to arbitrary thought, and published work indicates it can be defeated by an uncooperative subject. What is **established** is narrower and still remarkable — information previously available only as a private neural state can be mapped statistically into a machine-interpretable semantic representation. In August 2026 the [[wiki/Oden Institute|Oden Institute]] appointed **Blake Bordelon** jointly with UT's Department of Neuroscience around neural-population dynamics, memory, scaling laws, and brain-machine interfaces, formalizing the bridge between the institute's inverse-problem mathematics and the neural substrate. The observational gradient of the field is therefore complete in outline: **environment, body, behavior, association, movement, transaction, and — under narrow cooperative conditions — semantic interior**. ## The Epistemic Engine and the Substrate That Runs It What binds these layers is not a shared database but a shared mathematics, and that mathematics has a specific national-security origin. In 2008, UT Austin received approximately **$17 million** through NNSA's **[[wiki/Predictive Science Academic Alliance Program|Predictive Science Academic Alliance Program]]** to establish the **Center for Predictive Engineering and Computational Sciences**, PECOS, under a program created because the United States must sustain confidence in systems for which unrestricted full-scale nuclear explosive testing is unavailable. That constraint forced predictive science to answer the question machine-learning systems habitually postpone: **how much should the model be believed**. Verification asks whether the equations were solved correctly; validation asks whether the mathematical representation corresponds adequately to physical reality for its intended use; uncertainty quantification propagates measurement error, parameter uncertainty, model-form discrepancy, and numerical error into the prediction itself. PECOS institutionalized this as **decision-grade predictive computation**, and Oden researcher **Tan Bui-Thanh** subsequently received DTRA support to develop reduced-order models permitting real-time prediction of nuclear-weapon radiation effects, because high-fidelity calculation was too slow for timely decision — the same investigator later applying the machinery to real-time uncertainty quantification for infectious-disease outbreaks. The bridge can be stated in one line: **VVUQ determines how much the reconstructed state should be trusted; reduced-order modeling determines whether the state can be reconstructed fast enough to matter operationally.** A biosurveillance system asks whether its sources justify belief that an outbreak exists; a biometric system asks whether a presented identity is authentic; a sonar system asks what hidden object generated the observed acoustic field; an analyst asks what concealed activity best explains fragmentary reporting; a nuclear-effects model asks what physical state obtains under conditions that cannot be repeatedly reproduced. In every case: **partial observations → hidden-state estimate → quantified confidence → forecast → decision**. The computational substrate beneath that loop has changed order of magnitude. UT reported in late 2025 that it had accumulated more than **5,000 advanced NVIDIA GPUs**, and by July 2026 the 4,000-GPU Blackwell component of **Horizon** had begun operation, with the completed system designed around roughly one million CPU cores, approximately **20 exaflops of BF16/FP16 AI performance, up to 80 exaflops at FP4, and 400 petabytes of solid-state storage** in a new liquid-cooled Round Rock facility. The relevance is not raw scale. Uncertainty quantification is more expensive than deterministic simulation because it requires ensembles, repeated inversion against new data, sensitivity analysis, and continual updating, and a persistent digital twin multiplies the requirement again because the representation must stay synchronized with reality rather than terminate after producing a result. Horizon expands not how **large** a representation can be but how **computationally alive** it can remain — how many hypotheses can be held simultaneously, how quickly observations can be assimilated, and how thoroughly uncertainty can be characterized before an irreversible decision. ## IC ITE Removes the Requirement That the Brain Sit Beside the Sensor The final structural requirement is that the data, the model, and the decision-maker need not occupy the same organization or the same geography. The **Intelligence Community Information Technology Enterprise**, approved in 2011 and formally strategic from 2012, relocated intelligence integration out of bureaucratic negotiation and into the technical architecture itself through a common desktop environment, dual cloud provision, an applications marketplace, network engineering, enterprise management, and — decisively — **identification, authentication, and authorization services**. The IC's **Unified Identity Attribute Set**, published in the technical-standards environment from 2012, extends enterprise identity beyond human users to **non-person entities such as machines, servers, services, processes, and applications**, giving machine actors a formal position inside the authorization architecture. The 2017–2021 IC Data Strategy openly criticized dependence on bilateral sharing agreements and program-specific databases, calling instead for data to be freed from those dependencies, cataloged, self-described, discoverable by automated means, and governed by **machine-executable data-management rules**. The 2019 cloud strategy calls for secure access to functions, capabilities, and data anywhere, anytime, and under all conditions, including disconnected and edge operation, while supporting AI, machine learning, and high-performance computing across security fabrics. The operational precedent for that separation is **Saturn Arch**, an aerial IED-detection and intelligence capability originating under the National Geospatial-Intelligence Agency whose mission transferred in 2013–2014 to Army Intelligence and Security Command and **Task Force ODIN-East** — described by the Army as the first complete transfer of a mission from a national-level intelligence agency into the conventional Army — and which had already moved distributed processing, exploitation, and dissemination away from theater sensors into protected reachback facilities in the United States. The sensor stayed forward; the analytic brain did not. **Task Force ODIN — Observe, Detect, Identify, Neutralize** — compresses the whole control loop into four verbs, and read as an information architecture rather than a battlefield instruction each verb is a stage: observe is authorized acquisition; detect is discrimination of meaningful events from volume; identify is entity resolution, authentication, and contextual attribution; and **neutralize is conversion of intelligence into an authorized consequence, which need not be kinetic** — it can be warning, denial of access, credential revocation, interdiction, further collection, watchlisting, investigation, or escalation to a human decision authority. Army oral histories independently attest personnel describing heterogeneous platforms as operating **"under the ODIN umbrella,"** which matters because an umbrella survives conceptually even when its components remain distinct organizations, programs, and contractors. That is precisely the relationship this article describes in Austin: not one program, but a common computational grammar joining specialized institutions that retain their own authorities. ## The Oversight Record, Read Honestly in Both Directions A fair evaluation must ask whether formal program descriptions guarantee that every funded activity unfolded exactly as stated. They do not, which is why Inspectors General, protected-disclosure channels, and audits exist at all. DTRA maintains an Inspector General hotline through which military personnel, civilian employees, contractors, and private citizens can report unauthorized disclosures, reprisal, mismanagement, abuse, and procurement irregularities — including **separate classified reporting paths on SIPRNET and JWICS**. That detail cuts precisely against overconfident negative inference: the open record cannot establish whether classified complaints exist, and the absence of a public whistleblower is therefore evidence of nothing beyond the boundaries of what is publicly reportable. The Department of War Inspector General has separately issued **DOWIG-2026-085, _Audit of the Defense Threat Reduction Agency's Management of Grants_**, initiated in May 2025 to determine whether DTRA effectively managed grants in accordance with terms and conditions; the audit identified a reviewed universe of grants lacking adequate assurance of effective management owing to late or missing deliverables and incomplete grant documentation. The grants examined postdate Surety BioEvent substantially and the audit **does not implicate Barber, Meyers, Ellington, BioEvent, or the Austin biosurveillance lineage**. Its significance is institutional rather than accusatory: a grant abstract is not a comprehensive surveillance instrument over everything occurring inside a large research enterprise. The biological-counterterrorism infrastructure has produced exactly one strong public whistleblower, and his disclosure is narrower than it is often reported to be. **Harry Jackson**, a former DHS information-security manager associated with **[[wiki/BioWatch|BioWatch]]** — the nationwide environmental biological-attack detection system using air samplers in major metropolitan areas — raised concerns beginning in 2016–2017 that the BioWatch portal exposed sensitive information including sampler locations, detection results, and response procedures through insecure infrastructure, and alleged retaliation after demanding remediation; DHS OIG subsequently reported substantiating retaliation against a whistleblower who had complained about mishandling of classified BioWatch information. Jackson is **not evidence of covert observation of individual people**. He is evidence that insiders were raising serious concerns about secrecy, information handling, and oversight inside the biological-counterterrorism infrastructure, which is a different and still material fact. The resulting evidentiary hierarchy is clean and should be stated as such. **Loder** demonstrates an insider alleging unlawful retention and collection concerning innocent people inside the federally supported counterterrorism fusion architecture. **The Quiet Skies insiders** demonstrate covert, individualized physical surveillance of unwitting Americans generated by counterterrorism risk algorithms and association rules. **Jackson** demonstrates oversight failure and retaliation inside an actual biological-threat detection network. What does not exist in the public record is the final bridge: an insider stating that **DTRA-funded biosurveillance research was itself used to covertly observe unwitting human targets**. ## What Has Not Turned Up, Stated as Falsifiable Propositions Discipline requires naming the specific claims that would change the picture and reporting honestly that none of them is presently supported. No public whistleblower, Inspector General finding, litigation record, IRB disclosure, or grant document establishes that Surety BioEvent used named Austin residents as experimental subjects; that Barber, Meyers, or Ellington recruited watchlisted travelers; that ARL:UT or the Army UARC caused individuals to relocate to Austin for undisclosed experimentation; that Google Fiber infrastructure in the lower Colorado Street–Second Street corridor was connected to BioEvent; that PECOS models were applied to particular Austin residents; that ARIC selected people for a UT or DTRA research cohort; or that TSA or FBI watchlist status was knowingly used to route subjects into an Austin university experiment. Each is a **discriminating proposition** whose confirmation would materially alter the analysis. Each is presently **unsupported**. The absence of that bridge does not collapse the architecture. It changes what the architecture is, and the change is an escalation rather than a retreat. ## The Field Does Not Require a Door Marked Experiment Consider an ordinary person arriving in Austin during the mature period of this infrastructure. Before arrival, the reservation has entered Secure Flight, and if identifiers or circumstances trigger a rule the individual will not be told. The aircraft lands with passenger and itinerary information already resident at the federal level. A later police encounter can generate a routine database query against watchlist information, and a confirmed match can produce a fusion-center notification carrying location, time, vehicle, and biographical detail, followed by additional database queries at the regional node. ARIC occupies that regional layer, and its leaked historical material shows that lawful conduct could be reported through Threat Liaison Officers and retained after an analyst found no nexus — with the Austin records reachable, by design, from the North Texas analytical portal as well. Ordinary searches join an aggregate behavioral corpus from which DTRA-funded methods had already demonstrated that population disease state could be estimated. Movement through the downtown corridor coincides with a documented Google physical presence that has no established programmatic connection to that research. A hospital visit contributes to legally authorized public-health surveillance from which authorized departments may re-identify on alert. Residence in a monitored catchment contributes biological material to a population estimate resolvable toward buildings. A nearby extraordinary crime can place a phone inside a geofence dataset. A search history can satisfy a court-authorized keyword warrant and enter an investigative result set before investigators know the person's name. And a bystander designated as an intelligence liaison can convert an innocuous behavior into an intelligence record. None of these mechanisms requires the subject to comprehend that separate institutions are constructing partial representations of the same physical human being. More importantly, **none of them requires the institutions to share a database**. They need only render the person **machine-addressable from independent directions**, which is exactly what the IC's own architectural literature describes as the objective: heterogeneous systems remaining distinct while data, identity, and policy become interoperable and machine-readable. This is the true content of the phrase _a densely instrumented environment into which an unwitting individual can enter naturally_, and the word **unwitting** does not have to mean subjected to an undisclosed experiment. It can mean something far more ordinary and far more general. The individual knows they boarded an airplane but not how Secure Flight classified the itinerary. They know they drove through the city but not whether a routine stop queried a federal watchlist. They know they searched but not that aggregate search behavior contributes to an epidemiological state estimate. They know they entered an emergency department but not that the resulting record is a sensor reading. They know they flushed a toilet but not that the effluent may be interrogated for viral signature. They know they walked through a neighborhood but not whether the device in their pocket entered a geofence dataset. They know they spoke, prayed, protested, mailed a package, or posted online but not whether a designated liaison judged the behavior reportable. And Quiet Skies supplies the most extraordinary confirmation available: thousands of Americans **did not know that trained federal agents were physically following them and documenting their behavior minute by minute**. Transportation systems establish identity and movement. Fusion centers join local observation to federal intelligence and, in Texas, to each other across metropolitan areas. Public-health systems convert routine medicine into population sensing with lawful re-identification pathways. Environmental surveillance converts biological waste into epidemiological signal at progressively finer spatial resolution. Commercial platforms convert behavior into searchable corpora. Geofence and keyword methods reverse the direction of investigation, beginning with a place, a time, or a query and discovering identities afterward. University defense programs develop the sensors, trust metrics, prediction systems, and countermeasures. PSAAP and PECOS formalize how uncertainty itself becomes a computable engineering quantity. IARPA Odin hardens identity against deception while a prime performer sits in Austin. IC ITE supplies the enterprise in which computation, identity, policy, and data no longer need to remain beside the sensor. ODIN demonstrates persistent observation converted into identification and authorized consequence. DTRA carries the identical epistemic machine into the WMD domain, where **detection necessarily precedes attribution** because the first evidence of a deliberate biological event is clinically indistinguishable from an unusual natural one. The whistleblowers establish that the institutional boundary between **possible threat** and **closely observed person** has, in documented federal programs, expanded far enough to capture thousands of people never accused of anything. The Austin leaks establish that a local intelligence node accumulated and retained information about lawful conduct after its own analysts found no terrorism nexus. The warrant record establishes that investigators in this specific city sought unknown identities by querying the spatial and behavioral exhaust of entire user populations. The biosurveillance record establishes that human populations supply useful state information without individually enrolling in anything. These are not speculative capabilities awaiting demonstration. They executed. What remains **unresolved** is the narrow and specific proposition that someone arranged for a particular unwitting person to come to Austin so that these architectures could observe them. The available evidence suggests something structurally more consequential than that proposition would be if proven. By the time all of these systems existed simultaneously, **nobody needed to arrange it**. A person could arrive for entirely ordinary reasons and cross enough invisible informational boundaries that separate institutions could observe identity, movement, association, behavior, health signal, digital trace, and environmental emission without that person ever perceiving a single unified apparatus — because there is no single unified apparatus, and its absence is not a mitigation. A conspiracy has a perimeter, a membership, and a failure mode. A **field** has none of these. It has coverage. That is what makes Austin worth the study. Not a demonstrated undisclosed experiment. A **densely instrumented computational environment in which the subject becomes observable merely by living normally**, in which lawful authority, ordinary medicine, commercial convenience, municipal sanitation, and academic mathematics each contribute a facet of the same machine-readable person, and in which the operative question long ago stopped being _who is watching_ and became **what is addressable**. --- [[about/About Bryant McGill|Bryant McGill]] is a Wall Street Journal and USA Today bestselling author, systems architect, technologist, and strategic advisor, as well as a Congressionally Recognized Ambassador of Goodwill and United Nations–appointed Global Champion. His work spans naval intelligence systems, computational linguistics, artificial intelligence, digital transformation, and civilizational governance architecture. His forward analysis on U.S.–Israel Pax Silica frameworks has appeared in Jewish/Jerusalem News Syndicate (JNS). --- ## References 1. U.S. Senate Committee on Homeland Security and Governmental Affairs — [Examining the Weaponization of the Quiet Skies Program](https://www.hsgac.senate.gov/hearings/examining-the-weaponization-of-the-quiet-skies-program/), full-committee hearing, September 30, 2025, with testimony of Senior Federal Air Marshal Mark Crowder. 2. Congress.gov — [S. Hrg. 119-215, Examining the Weaponization of the Quiet Skies Program](https://www.congress.gov/event/119th-congress/senate-event/LC75186/text), hearing transcript including Crowder's opening statement and biography. 3. _The Boston Globe_ — [In "Quiet Skies," TSA Is Tracking Regular Travelers Like Terrorists in Secret Surveillance](https://apps.bostonglobe.com/news/nation/graphics/2018/07/tsa-quiet-skies/), July 2018, the original air-marshal disclosure. 4. U.S. Government Accountability Office — [Terrorist Watchlist: FBI Should Improve Outreach Efforts to Nonfederal Users, GAO-26-108650](https://www.gao.gov/products/gao-26-108650), published January 12, 2026; public version of a sensitive report issued August 2025. 5. U.S. Government Accountability Office — [GAO-26-108650 full report PDF](https://www.gao.gov/assets/gao-26-108650.pdf), including Appendix II on the role of fusion centers in the watchlist encounter process. 6. U.S. District Court for the District of Maine — [Loder v. Maine State Police, No. 2:20-cv-00157, Document 74](https://www.govinfo.gov/content/pkg/USCOURTS-med-2_20-cv-00157/pdf/USCOURTS-med-2_20-cv-00157-2.pdf), the judicial record underlying the December 2022 verdict. 7. _Portland Press Herald_ — [Jury Awards $300,000 to Retired Trooper Who Sued State Police for Retaliation](https://www.pressherald.com/2022/12/02/jury-deliberations-begin-in-retaliation-trial-brought-by-retired-state-trooper/), December 2, 2022. 8. DHS Office of Inspector General — [Semiannual Report to Congress, October 2022–March 2023](https://www.oig.dhs.gov/sites/default/files/assets/SAR/2023/oig-sar-oct22-mar23.pdf), containing the substantiated BioWatch-related whistleblower retaliation finding. 9. Defense Threat Reduction Agency — [DTRA IG Hotline](https://www.dtra.mil/Contact-Us/DTRA-IG-Hotline/), including classified reporting paths on SIPRNET and JWICS. 10. Department of War Office of Inspector General — [Audit of the Defense Threat Reduction Agency's Management of Grants, DOWIG-2026-085](https://www.oversight.gov/inspectors-general/department-war-oig), and the [May 19, 2025 project announcement memorandum](https://media.defense.gov/2025/May/21/2003720006/-1/-1/1/D2025-D000AU-0114.000.PDF). 11. Defense Threat Reduction Agency — [The Biosurveillance Ecosystem (BSVE) Fact Sheet](https://www.dtra.mil/Portals/61/Documents/CB/BSVE%20Fact%20Sheet_04282015_PA%20Cleared.pdf), describing the AWS-based unclassified fusion environment and its SDK-extensible analytics. 12. U.S. Department of Defense — [DTRA Scientists Develop Cloud-Based Biosurveillance Ecosystem](https://www.defense.gov/News/News-Stories/Article/Article/681832/dtra-scientists-develop-cloud-based-biosurveillance-ecosystem/), on aggregation of open-source, social-media, diagnostic, DoD, interagency, and international surveillance data. 13. Defense Threat Reduction Agency — [The Biological Threat Reduction Program](https://www.dtra.mil/Portals/125/Documents/CTR-Factsheets/DTRA-Fact-Sheet-BTRP-Aug-2023.pdf), joining biosafety, biosecurity, biosurveillance, outbreak detection, and nonproliferation in a single mission. 14. Texas ECE, UT Austin — [Prof. Suzanne Barber Awarded DTRA Grant for Work on Surety BioEvent App](https://ece.utexas.edu/news/prof-suzanne-barber-awarded-dtra-grant-work-surety-bioevent-app), the announcement describing the six-dimensional source-trust representation. 15. UT News — [Flu Season Forecasts Could Be More Accurate with Access to Health Care Companies' Data](https://news.utexas.edu/2018/09/19/this-data-source-could-enable-better-flu-forecasts/), on the evaluation of 600-plus data streams and transfer of methods into DTRA's Biosurveillance Ecosystem. 16. _Forbes_ — [Government Secretly Orders Google to Identify Anyone Who Searched a Sexual Assault Victim's Name, Address or Telephone Number](https://www.forbes.com/sites/thomasbrewster/2021/10/04/google-keyword-warrants-give-us-government-data-on-search-users/), including EFF's identification of the additional Austin bombing keyword warrants. 17. _Northwestern University Law Review_ — [Moderating Reverse Internet Keyword Warrants](https://scholarlycommons.law.northwestern.edu/cgi/viewcontent.cgi?article=1631&context=nulr), on the Google, Yahoo, and Bing orders in the Austin investigation. 18. _University of Pennsylvania Journal of Constitutional Law_ — [Mass Data Search Warrants and the Constitution](https://scholarship.law.upenn.edu/cgi/viewcontent.cgi?article=1903&context=jcl), citing the March 14, 2018 Austin Division warrant applications 18-MJ-168, 18-MJ-170, and 18-MJ-171. 19. City of Austin — [LEAP Interlocal Agreement with the North Central Texas Council of Governments](https://services.austintexas.gov/edims/document.cfm?id=176641), September 2012. 20. City of Austin — [Amended and Restated Interlocal Agreement for the Austin Regional Intelligence Center](https://services.austintexas.gov/edims/document.cfm?id=427414), recording the 2010 original agreement, the 2012–2013 Sustainment Funding Agreement, the **2017 Partner Equivalent Agency amendment** opening membership to non-governmental agencies, the 2022 partner additions, and the center's collection limitations, Privacy Officer, and Public Safety Commission audit path. 21. Department of Homeland Security — [Law Enforcement Information Sharing (LEIS) Service briefing](https://info.publicintelligence.net/DHS-LEISS.pdf), listing **LEAP (NCTCOG)** among established federal information-sharing partners alongside ARJIS, LInX, OneDOJ, and TDEx. 22. GlobeNewswire — [COPsync Announces Strategic Partnership With the Law Enforcement Analysis Portal (LEAP)](https://www.globenewswire.com/news-release/2010/01/12/412090/181727/en/COPsync-Announces-Strategic-Partnership-With-the-Law-Enforcement-Analysis-Portal-LEAP.html), January 12, 2010, establishing LEAP's origin date, its 200-agency and 75-county reach, and its ONEDOJ and ICEPIC access. 23. _The Trace_ — [How a Silicon Valley Start-up Is Helping Police Get Smarter About Solving Gun Crimes](https://www.thetrace.org/2019/10/police-data-google-for-gun-crime-forensic-logic-oakland/), October 2019, on Forensic Logic's LEAP, the Oakland deployment, and the 70,000 monthly queries. 24. Dallas Police Department — [MOSAIC: Metro Operations Support and Analytical Intelligence Center, Public Safety Update](https://www3.dallascityhall.com/committee_briefings/briefings0108/PS_010708_MOSAIC.pdf), January 2008. 25. SoundThinking — [COPLINK X Is Now CrimeTracer](https://www.soundthinking.com/blog/coplink-x-is-now-crimetracer/), the company's own account of the LEAP Network, the 2017 acquisition of COPLINK Detect from IBM, the 2022 ShotSpotter acquisition, and the rebrand. 26. SoundThinking — [CrimeTracer](https://www.shotspotter.com/law-enforcement/coplink-x/), advertising more than one billion records across thirty-plus data sources including RMS, CAD, license plate readers, jail management, court, and warrant records. 27. ShotSpotter, Inc. — [Annual Report on Form 10-K, March 14, 2023](https://fintel.io/doc/sec-shotspotter-inc-1351636-10k-2023-march-14-19430-4719), recording the January 2022 acquisition of Forensic Logic and the integration of COPLINK X into the platform. 28. IARPA — [Odin](https://www.iarpa.gov/research-programs/odin), including the Thor and Loki structure, performance targets, and the four prime performers. 29. U.S. Army — [Saturn Arch](https://www.army.mil/article/119068/saturn_arch), on the transfer of a national-agency mission to INSCOM and Task Force ODIN-East. 30. U.S. Central Command — [Task Force ODIN Transfer of Authority](https://www.centcom.mil/MEDIA/NEWS-ARTICLES/News-Article-View/Article/1131834/task-force-odin-transfer-of-authority/), March 2017. 31. ODNI — [Intelligence Community Information Environment Data Strategy 2017–2021](https://www.odni.gov/files/documents/CIO/Data-Strategy_2017-2021_Final.pdf), on freeing data from bilateral agreements and machine-executable data-management rules. 32. ODNI — [IC Technical Specification: Unified Identity Attribute Set](https://www.odni.gov/files/documents/CIO/ICEA/IC_Tech_Spec_Attributes_V2_Final_PUBLIC.pdf), extending enterprise identity attributes to non-person entities. 33. ODNI — [The IC Information Environment Roadmap](https://www.odni.gov/files/documents/CIO/IC-IT-Roadmap-Vision-For-the-IC-Info-Environment-May2024.pdf), May 2024, on multi-cloud, data-centricity, Zero Trust, and tactical-edge operation. 34. NNSA — [Selection of the Next Round of Predictive Science Academic Alliance Program Centers](https://www.energy.gov/nnsa/articles/nnsa-announces-selection-next-round-predictive-science-academic-alliance-program), connecting PSAAP to verification, validation, uncertainty quantification, and stockpile certification. 35. Oden Institute — [Minimizing Uncertainty in the Uncertain World of Defense, Energy](https://oden.utexas.edu/news-and-events/news/minimizing-uncertainty-in-uncertain-world-of-defense-energy/), on Tan Bui-Thanh's DTRA-supported reduced-order models for nuclear-weapon radiation effects. 36. Oden Institute — [Predictive Engineering and Computational Sciences](https://oden.utexas.edu/research/centers-and-groups/predictive-engineering-and-computational-sciences/), the PECOS statement of purpose. 37. Oden Institute — [Collaborative Opportunities with Applied Research Laboratories](https://oden.utexas.edu/news-and-events/events/1305/), the October 2018 seminar documenting the ARL:UT–Oden interface. 38. UT News — [DOD Awards $1.1 Billion Contract to UT Austin's Applied Research Laboratories](https://news.utexas.edu/2017/09/28/dod-awards-11-billion-to-applied-research-laboratories/), September 28, 2017. 39. UT News — [UT's Lab That Engineers the Hidden Technologies Behind National Security and Everyday Life](https://news.utexas.edu/2026/07/24/uts-lab-that-engineers-the-hidden-technologies-behind-national-security-and-everyday-life/), July 24, 2026. 40. TACC Documentation — [Horizon System Overview](https://docs.tacc.utexas.edu/hpc/horizon/system/), on the Blackwell deployment, exaflop-class AI performance, and storage capacity. 41. UT News — [Brain Activity Decoder Can Reveal Stories in People's Minds](https://news.utexas.edu/2023/05/01/brain-activity-decoder-can-reveal-stories-in-peoples-minds/), the 2023 noninvasive semantic decoder and its cooperative-subject constraints. 42. Oden Institute — [New Faculty Blake Bordelon Joins UT Austin With a Vision to Connect Neuroscience and Machine Learning](https://www.oden.utexas.edu/news-and-events/news/Blake-Bordelon-UT-Austin-faculty-Neuroscience-and-Machine-Learning/), August 2026. 43. National Academies Press — [BioWatch and Public Health Surveillance: Evaluating Systems for the Early Detection of Biological Threats](https://nap.nationalacademies.org/read/12688/chapter/7), on surveillance capacities common to natural emergence and bioterrorism. --- **Concept cluster.** [[wiki/Biosurveillance|Biosurveillance]] · [[wiki/Syndromic Surveillance|Syndromic Surveillance]] · [[wiki/Wastewater Surveillance|Wastewater Surveillance]] · [[wiki/Environmental Pathogen Surveillance|Environmental Pathogen Surveillance]] · [[wiki/Source Trust Tuple|Source Trust Tuple]] · [[wiki/Verification Validation and Uncertainty Quantification|Verification Validation and Uncertainty Quantification]] · [[wiki/Detection Before Attribution|Detection Before Attribution]] · [[wiki/Sensor-to-Action Loop|Sensor-to-Action Loop]] · [[wiki/Machine-Executable Policy|Machine-Executable Policy]] · [[wiki/Attribute-Based Access Control|Attribute-Based Access Control]] · [[wiki/Biometric Presentation Attack Detection|Biometric Presentation Attack Detection]] · [[wiki/Inverse Problem|Inverse Problem]] · [[wiki/Semantic Neural Decoding|Semantic Neural Decoding]] · [[wiki/Multi-Use Inference and Control|Multi-Use Inference and Control]] **Laboratory and institutional cluster.** [[wiki/Applied Research Laboratories at UT Austin|Applied Research Laboratories at UT Austin]] · [[wiki/Institute for Advanced Technology|Institute for Advanced Technology]] · [[wiki/PECOS|PECOS]] · [[wiki/Oden Institute|Oden Institute]] · [[wiki/Texas Advanced Computing Center|Texas Advanced Computing Center]] · [[wiki/Horizon|Horizon]] · [[wiki/Austin Regional Intelligence Center|Austin Regional Intelligence Center]] · [[wiki/Law Enforcement Analysis Portal|Law Enforcement Analysis Portal]] · [[wiki/Dallas MOSAIC|Dallas MOSAIC]] · [[wiki/Texas Fusion Center|Texas Fusion Center]] · [[wiki/Defense Threat Reduction Agency|Defense Threat Reduction Agency]] · [[wiki/Biosurveillance Ecosystem|Biosurveillance Ecosystem]] · [[wiki/Surety BioEvent App|Surety BioEvent App]] · [[wiki/IARPA Odin|IARPA Odin]] · [[wiki/IARPA MOSAIC|IARPA MOSAIC]] · [[wiki/Task Force ODIN|Task Force ODIN]] · [[wiki/Saturn Arch|Saturn Arch]] · [[wiki/IC ITE|IC ITE]] · [[wiki/HID Global|HID Global]] · [[wiki/Whole Communities–Whole Health|Whole Communities–Whole Health]] · [[wiki/Google Flu Trends|Google Flu Trends]] <!-- BEGIN AUSTIN SURVEILLANCE FIELD --> ## Wiki integration **Master:** [[research/The Austin Executable Loop|The Austin Executable Loop]] · **Branch map:** [[wiki/Austin Surveillance Field|Austin Surveillance Field]] · **Evidence:** [[wiki/Austin Surveillance Evidence Ledger|Austin Surveillance Evidence Ledger]] ### LEAP and the federal interface [[wiki/Law Enforcement Analysis Portal|LEAP]] · [[wiki/North Central Texas Council of Governments|NCTCOG]] · [[wiki/DHS Law Enforcement Information Sharing Service|DHS LEIS]] · [[wiki/ICEPIC|ICEPIC]] · [[wiki/Interlocal Information Sharing|Interlocal Information Sharing]] · [[wiki/LEAP Provider Transition|LEAP Provider Transition]] The integration checked the **September 27, 2012 minutes, items 57–58, each approved 7–0**, and the DHS briefing’s **existing-MOU** listing. Additional NCTCOG records locate its contract closeout on **August 31, 2014**, with transition to new providers whose identities remain to be established. The city agreement also records 2005 project approval, 2006 advisory-committee approval, and automatic annual renewal after its initial term. These newly checked details are developed in the wiki without replacing the essay’s source prose. ### Full-field routes [[wiki/Austin Regional Intelligence Center|ARIC]] · [[wiki/Partner Equivalent Agency Agreements|Partner Equivalent Agency Agreements]] · [[wiki/Watchlist Encounter Process|Watchlist Encounter Process]] · [[wiki/Quiet Skies|Quiet Skies]] · [[wiki/Reverse Search Warrants|Reverse Search Warrants]] · [[wiki/Population State Estimation|Population State Estimation]] · [[wiki/Observation and Human-Subject Research|Observation and Human-Subject Research]] · [[wiki/Fusion Center Oversight|Fusion Center Oversight]] · [[wiki/Surveillance Record Persistence|Surveillance Record Persistence]] [[wiki/Oden Institute|Oden]] · [[wiki/PECOS|PECOS]] · [[wiki/Suzanne Barber|Barber]] · [[wiki/Sensor and Social Networks Armed to Detect and Defend against Terrorist Attacks|DHS 2009 sensor/social-network project]] · [[wiki/Surety BioEvent App|BioEvent]] · [[wiki/Computational Addressability|Computational Addressability]] · [[wiki/Multi-Use Inference and Control|Multi-Use Inference and Control]] <!-- END AUSTIN SURVEILLANCE FIELD -->