Part of a series on |
Global surveillance |
---|
Disclosures |
Systems |
Selected agencies |
Places |
Laws |
Proposed changes |
Concepts |
Related topics |
Sentient is a heavily classified artificial intelligence satellite intelligence analysis system of the United States Intelligence Community, operated by the National Reconnaissance Office (NRO) and developed by their Advanced Systems and Technology Directorate (AS&T), with the United States Air Forces Research Laboratory at Wright-Patterson Air Force Base and the Department of Energy's National Laboratories.[1][2][3][4][5]
Available information describes it as a complex automated system that allows intelligence agencies of the United States and the United States Armed Forces to use artificial satellites in Earth orbits to track in real time any objects detected or photographed, and to automatically repurpose with artificial intelligence and machine learning the tracking of targets, and to even decide which targets are worth tracking.[2][3][1][4]
Known public records of Sentients development programs and process date from 2009-2010 onward.[6][4][7] NRO emails from 2021 disclosed that the Unidentified Aerial Phenomena Task Force (UAPTF), which later became the All-domain Anomaly Resolution Office (AARO), was involved with the NRO and the Sentient program.[8]
A later declassified May 2009 report to the United States Congress, "FY 2010 Congressional Budget Justification, Volume IV," contains details about the National Reconnaissance Offices plans for real-time and updated satellite signals intelligence, providing context on NROs space-based missions and programs to collect data, such as Sentient, which would initially begin soliciting defense and related industry feedback in 2010.[9][6]
The US is arguably more reliant on overhead collection than ever before. To a large extent, satellite reconnaissance is the foundation for global situational awareness, and as such, it is an essential underpinning of the entire US intelligence effort. Space collection provides unique access to otherwise denied areas to provide persistent and responsive collection; and it does so without risk to human collectors or infringing upon the territorial sovereignty of other nations. It is also enables users to quickly focus on almost any point of the globe to rapidly respond to emerging situations or to monitor ongoing events. The NRO provides direct support to the war on terrorism, deployed military forces, and other IC and DoD activities requiring near real-time situational awareness and sustained high resolution/high sensitivity collection capability 24-hours-a-day, 7-days-per-week.
The NRO continued to explain to Congress the AR&D activities (research and development) goals of its Advanced Systems and Technology Directorate related to GEOINT, SIGINT, and Multi-INT collection of intelligence data.[10] Their areas of focus were detailed as monitoring any target across wide areas on-demand, including new phenomenological models for detecting non-traditional observables from targets of interest; geolocating them; increasing performance for collection against weak signals and low reflectance targets in dense clutter and co-channel interference backgrounds; and improvements in Phased Array technologies.[10]
The earliest known public references to Sentient date to 2010, when the NRO put forward a public request for information (RFI) requesting white papers on the concepts that would become Sentient, including user interaction technology, computing self-awareness and cognitive processing, and automated system processing.[1][9] Official NRO documents from 2012, declassified in 2019, describe it as "an on-going Research and Development (R&D) program, which is managed and operates out of the National Reconnaissance Office (NRO)".[4] NRO documents detail the Sentient program was in some form of initial development from 2013 through 2016.[11]
In 2019, a declassified National Reconnaissance Office document detailed Sentient as:[7]
Sentient collects vast amounts of complex information, buried in noisy data, and extrapolates the relevant pieces. What once required teams of analysts now uses automation to speed the information flow, enabling quicker, well-informed decisions. Analysts can focus on situational understanding instead of sifting through data-working smarter, not harder. Sentient will continue to investigate various alerting schemes and tasking methodologies for improving overall employment of NTM imaging capabilities. Through predictive analytics and machine learning, Sentient will increase the efficiency and effectiveness of NTM collections.
The program architecture was developed to demonstrate advanced technologies and techniques to revolutionize the then current intelligence analysis process across the Intelligence Community (IC).[4][12] Sentient utilizes "tipping and queueing", described as a method to coordinate multiple discrete sensor systems to optimize real-time monitoring over large areas from multiple systems; coordinating in effect multiple satellites to hand off and coordinate tracking of targets as the target leaves the area of one set of satellites and enters new areas.[13][14][15] A principle purpose of the Sentient system is described by the NRO as compiling at machine, versus human speed, synthesis of complex distributed data sources for rapid analysis faster than humans can manage.[16][8]
The Rand Corporation in 2016 summarized a major benefit of Sentient as allowing humans performing intelligence analysis to focus on the "so what?" aspect of analysis, while the Sentient automation would handle the details of "what", as automation may be better suited for the data collection role.[13] Writing for Grey Dynamics and analyzing NRO declassified collateral, Alec Smith noted Sentient's ability to increase "situational awareness based on observed activity and historical intelligence to model and anticipate potential courses of actions of adversaries".[17]
According to Robert Cardillo, former Director of the National Geospatial-Intelligence Agency, the system is intended to use "automated inferencing" to aid intelligence collection.[18] Cardillo summarized Sentient in 2017 as "...a research and development effort, conducted jointly with NRO, to experiment with automation that ingests data, makes sense of it in the context of an intelligence problem, and then infers likely future intelligence and collection needs... The approach is to continue to introduce automation in the processing arena that will support automated inferencing, and therefore, faster tasking for future collections."[18][17]
United States Army Captain Anjanay Kumar proposed in 2021 that should the United States Joint Forces Command engage in conflict with China over Taiwan, Sentient would be a likely target of Chinese forces.[19] Kumar described the scenario as, "While the Sentient program itself may be secure, the contractors and infrastructure associated with its hardware are not and present a weak point for the PRC to exploit. Before the data moves from the servers, malicious software will attack the servers and causes their physical destruction."[19]
The Verge described Sentient as “an omnivorous analysis tool, capable of devouring data of all sorts, making sense of the past and present, anticipating the future, and pointing satellites toward what it determines will be the most interesting parts of that future.”[1] Quoted in The Verge, Steven Aftergood of the Federation of American Scientists and the Project on Government Secrecy, said of Sentient, "It could include electronic intercepts of international communications; it could include prior imagery; it could include human sources."[1] In the same report, retired Central Intelligence Agency analyst Allen Thomson stated, "As I understand it, the intended — and aspirational — answer is ‘everything,’" and said that in "addition to images, that could include financial data, weather information, shipping stats, information from Google searches, records of pharmaceutical purchases, and more."[1] In The Fragile Dictator: Counterintelligence Pathologies in Authoritarian States by Wege and Mobley, they compared the NRO's Sentient program to the commercial Blacksky Global geospatial intelligence service offered by Spaceflight Industries, and noted that Sentient-type tools can increase "intelligence equities" in regard to oceanic shipping and "sanctions busting" attempts by authoritarian nation-states.[20] Sarah Shoker noted in Military-Age Males in Counterinsurgency and Drone Warfare that Sentient and comparable systems like automatic target recognition (ATR) can remove human limitations from time-sensitive intelligence analysis, highlighting that past behavioral and tactical patterns interpreted by these systems can forecast similar future actions.[21]
The National Reconnaissance Office revealed in 2022 that the Sentient program had involvement with the United States Department of Defenses Unidentified Aerial Phenomena Task Force (UAPTF) in tracking and research of unidentified aerial phenomena, also known as unidentified anomalous phenomena or unidentified flying objects.[8] In June 2022, the NRO declassified and released two unique and redacted reports on Sentient activities related to a UAP sighting of what was described as a "tic tac" shape, and other successes at detecting UAP.[2][3] NRO Office of Public Affairs official Karen Furgerson reported to The Verge in 2019 that:[1]
"The NRO has not said much about Sentient publicly because it is a classified program and NRO rarely appears before Congress in open hearings. It ingests high volumes of data and processes it. Sentient catalogs normal patterns, detects anomalies, and helps forecast and model adversaries’ potential courses of action. Sentient is a thinking system. Sentient aims to help analysts ‘connect the dots’ in a large volume of data."
In his book The Origins of Victory: How Disruptive Military Innovation Determines the Fates of Great Powers, Andrew Krepinevich highlights the same remarks from Furgerson, and notes that the "avalanche" of data available from intelligence, military, and commercial sources would overwhelm human staff attempting to analyze it all.[22] Krepinevich provides details on some of the commercial sources of data the NRO has contracted with to provide intelligence services and data for the Sentient program.[22] Among these are Maxar Technologies, Planet, and BlackSky; Krepinevich notes that BlackSky "hoovers up" materials from dozens of satellites, over one hundred million mobile devices, and similar numbers of ships, planes, social networks, and environmental sensors as raw collateral for big-data analytics to assist Sentient and it's machine learning operations.[22]
Betty J. Sapp, former Director of the National Reconnaissance Office from 2012 to 2019, described the Sentient program to the United States House Armed Services Subcommittee on Strategic Forces as "...a 'thinking' system that allows automated, multi-intelligence tipping and cueing at machine speeds-has been focused against many of our most challenging mission sets, resulting in new intelligence information that human-in-the-loop systems would have missed. Our Future Ground Architecture will leverage Sentient, and create an integrated cloud-based enterprise that will share tasking and intelligence products quickly across each of our ground sites, increasing both performance and resilience.[14] At the 2013 GEOINT Symposium held by the United States Geospatial Intelligence Foundation, Sapp said of Sentient, "The ground needs to be a very capable orchestrator to realize the full potential of multi-INT overhead space architecture," and that Sentient will 'will allow NRO to be not only responsive, but also predictive, with where it aims spaceborne assets.'[23] In 2016, Sapp discussed the application of using cubesat systems for NRO work, which were tested as part of the Sentient program.[24]
According to Sapp, the NRO has been asked to give more demonstrations to observe Sentient and its capabilities in demonstrations to the defense and intelligence communities than 'any other capability since the beginning of the organization’s history.'[25]
Declassified government documents have detailed that United States Government stakeholders involved in Sentient include the:[12][5][8]
This article incorporates public domain material from websites or documents of the United States Government.
{{cite web}}
: CS1 maint: multiple names: authors list (link)