AI War Cloud
AI decision making from battlefield to desktop
Keywords
artificial intelligence, artistic research, automated decision making, autonomous weapons, critical AI, defense, generative AI, international security, machine learning, military, surveillance, technoimperialism, tech industry, warfare
AI War Cloud is an interactive database and tangible interactive installation connecting and explaining the current techno-imperial boomerang that is perpetuated by machine learning.
/Users/thatmachine/miniconda3/envs/quart/lib/python3.14/site-packages/gravis/_internal/plotting/template_system.py:5: UserWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html. The pkg_resources package is slated for removal as early as 2025-11-30. Refrain from using this package or pin to Setuptools<81.
import pkg_resources as _pkg_resources
Details for selected element
When the bots, recommender systems, and automated agents many people use daily are the same technologies used to wage war, how should we understand and hold these systems to account? Ultimately, what responsibilities do tech makers and users have in choosing AI tools, when their development also leads to deadly outcomes at massive scales? With the spotlight now on systems like Ukraine’s Palantir MetaConstellation and Israel’s Lavender, the stakes for machine learning tasks are increasingly urgent and personal. This project examines the specific machine learning tasks used in military “AI Decision Support Systems” (AI-DSS). These combine massive amounts of data and processing to help make choices about who lives or dies, speeding up the process exponentially. The research presents a database that details how the training datasets, models, and inferences military tools rely on are the very same types used by consumers. It shows how these are also deployed by, or even upon, citizens of the countries that first developed them, after they are tested on vulnerable foreign populations in conflict zones.
Loading ITables v2.9.1 from the init_notebook_mode cell...
(need help?)
|
| ⓘWeapon | Developed | Used By | Military Purpose | Type of Tech | Repurpose (Potential/Actual) | SourceType |
|---|---|---|---|---|---|---|
| Lattice | 2025 | US, Anduril | "Lattice accelerates complex kill chains by orchestrating machine-to-machine tasks at scales and speeds beyond human capacity"; "provides a way to link up and control different robots, sensors and other military equipment, a kind of operating system for war"; combines with company's sensor towers, drones, cruise missiles | VR interface, computer vision, edge computing, natural language processing, autonomous data processing and analysis, "detect, track, and classify" | US Customs and Border Patrol | Corporate, Press |
| Large Geospatial Models | 2025 | Niantic Labs | (proposed) | Large generative model combined with 3D photogrammetry of public spaces for enhanced/predictive mapping capacities | Pokémon Go | Press |
| Polygraph+ | 2024 | US, Presage, Altec | Credibility Assessment, updated polygraph, physiological tracking, biofeedback, AI analysis | "standard consumer grade cameras for medical assessment" "research and clinical technologies that monitor biosignals" | Interrogation, job interviews, "vetting new hires, evaluating existing DoD personnel for access to classified information, assisting in determining source credibility, and supporting criminal investigations" | Government |
| Lavender | 2023 | Israel | Creates kill lists. Ranks each person from 1-100 for their similarity to pre-existing targets, based on data found in WhatsApp groups, mobile phones, location targeting, network surveillance, face recognition, and other tracking techniques. collects "Visual information, cellular information, social media connections, battlefield information, phone contacts, photos. [...and] over time the machine will come to identify features on its own" | Ranking, classification, prediction, text analysis, image analysis | Works on similar principles to recommender systems: If you liked this reel, or behaved this way online, you are most like these other internet/mobile users. Uses the hypersurveillance of vast cyber infrastructure to connect information that may not represent "likeness." Also entangled with text/image analysis | Press |
| Where's Daddy | 2023 | Israel | Determines when people are at home with their families in order to destroy them and their homes | Geolocation, computer vision (object detection), signal tracking | Package delivery, advertising, legal preceedings | Press |
| Legion X | 2022 | Israel, Elbit Systems | Platform for identifying and targeting people carrying weapons or wearing uniforms, for controlling autonomous drone swarms, "coordinated strike after human confirmation" | Network, object identification, geolocation, weapon, pose estimation | Security, identify people in crowds, in shops | Corporate |
| Artificial Intelligence Platform (AIP) | 2022 | Palantir | "Artificial Intelligence Platform, which allows for the integration of large language models into the company's military products" "An Army of AI Agents that do what I say for me" | Large language model, natural language processing, + | Supply chain management, contracts, infrastructure grid, "LMNotebook" dashboards | Corporate, Tenders, Video |
| MetaConstellation | 2022 | Ukraine, Palantir, USA | "relies on intelligence gathered on enemy troop positions by commercial satellites, heat sensors and reconnaissance drones as well as spies working behind enemy lines and ordinary Ukrainians pin-pointing the locations of Russian troops on the country’s E-Enemy app. The software uses AI to transform the data into a map highlighting the probable locations of Russian artillery, tanks and troops. A Ukrainian soldier using a tablet device is given a list of coordinates and can then direct their fire. The technology also “learns” from previous strikes, meaning that it is constantly getting better at identifying and locating material." | Prediction, object detection, "process imagery, detect, and geolocate objects, and determine any movement" | "monitor locations across the globe for indications of competitor activities" | Corporate, Video |
| Foundry | 2022 | Palantir, USA | Data tools likely supporting Gotham and Palantir's other defense products | "industry agnostic" tools to create data structures and algorithms for combining and managing large amounts of information, with tools to perform data analysis and create applications. | Contracted by US Immigration and Customs Enforcement and other US government agencies to track migration and financial records, and also by UK NHS agencies to track pandemic health records | Academic, Press |
| Clearview | 2022 | Ukraine, ClearviewAI | Facial recognition to "identify dead soldiers and to uncover Russian assailants and combat misinformation." | Facial recognition | Commercial/policing tool then used for warfare | Corporate, Video |
| Wolly | 2022 | Ukraine, Roboneers, D3 (former Google employees, investors), Helsing (investors) | Automated machine guns “We could sit in the trench drinking coffee and smoking cigarettes and shoot at the Russians” | Object identification, computer vision, repurposed video game controllers | Security systems, prisons, protests | Press |
| First-Person-View Drones | 2022 | Ukraine, Vyriy, PG Robotics, Skyeton, Swarmer, others | Fully automated, weaponized drones, with automated targeting. Some can easily be built with hobby kit like materials. | Quads, computer vision, deep learning, classification and sorting algorithms | Personal drones, sports, already in use. Drones as local police first responders | Press |
| MAPLE | 2022 | United Kingdom | Command and control platform for the UK Navy to control an automated fleet of unmanned air, ground, and water vehicles | Assign, task | Driver and delivery apps, logistics apps | Corporate, Press |
| Mantacus | 2022 | Mantacus, DIANA (NATO) | AI software continuously scans and tracks people on CCTV for possible concealed weapons, using radar signals. | Signal analysis, object detection, classification | Private security, crowd control, covert surveillance | Corporate |
| The Alchemist | 2021 | Israel | Sends real-time data to commanders in field about possible threats, using machine learning | Transmit, analyse | As-yet unknown | Press |
| Project Nimbus | 2021 | Israel, Google, Amazon | Provides cloud services (storage, compute) "Nimbus is a flagship project and a key anchor in the implementation of the Israeli Government’s cloud policy, and is intended to provide a comprehensive and in-depth solution to the provision of public cloud services to the Government, the defense establishment and other public organizations. [...] accelerating the Government’s digital transformation process, as well as improving the command, control and cyber defense capabilities of the government IT systems. " | Cloud compute and storage, edge computing | Cloud compute and storage, edge computing | Press, Tenders |
| Assault Rifle Combat Application System | 2021 | Israel, Elbit Systems | a gunsight sold by the Israeli defense firm Elbit Systems. According to a company spec sheet, the “AI-powered” device is capable of “human target detection” at a range of more than 600 yards, and human target “identification” (presumably, discerning whether a person is someone who could be shot) at about the length of a football field. Anna Ahronheim-Cohen, a spokesperson for the company, told MIT Technology Review, “The system has already been tested in real-time scenarios by fighting infantry soldiers.” | Pose estimation, object detection, AR | Video games: “We made it very intuitive so it looks like PlayStation’s Fortnite ; it shows range, wind and ammo left, etc.,” said Arie Chernobrov, general manager of Elbit Security Systems. | Press |
| Lethal autonomous weapons (LAW) | 2021 | Libya, Azerbaijan (US, Turkey, Israeli-made), Russia (aka Kub & Lancet) | "“loitering munitions” — drones that can autonomously patrol an area and automatically divebomb enemy radar signals. These weapons look like smaller versions of the remote-controlled drones that have been used extensively by the U.S. military in Iraq, Afghanistan and other conflicts. Instead of launching missiles through remote control, though, loitering munitions have a built-in explosive and destroy themselves on impact with their target." | Automated drones, hardware, geolocation | Consumer drones | Government, Press |
| Gotham | 2021 | USA, Palantir | An "AI-powered kill chain," this operating system identifies targets and pairs them with weapons. Its command center supports mixed reality and edge computing. | Sorting, identifying, data analysis, edge computing | Miscellaneous business to business applications, "global project management" | Corporate |
| Joint Warfighting Cloud Capability, JEDI | 2021 | USA, Amazon, Microsoft, Google, Oracle, IBM | Multi-cloud provider for military use, including Unclassified, Secret and Top Secret (2021, 2019) | Storage, compute (calculation), "Commercial Cloud Enterprise" | Cloud compute and storage, edge computing | Government, Press |
| AI Verse | 2021 | DIANA (NATO), AI Verse | Creates synthetic (generated) image datasets used for training AI to detect people, objects, and weapons | NaN | NaN | Corporate, Government |
| The Gospel (Habsora) | 2021 | Israel | Marks homes and public buildings as targets for destruction, for alleged military use, based on "enormous amounts of data that “tens of thousands of intelligence officers could not process,” and recommends bombing sites in real time. Possible data includes: "cell phone messages, satellite images, drone footage, and even seismic sensors." | Classification, data analysis, object recognition | Redlining, unequal or biased distribution of resources | Press |
| (Mobile) Fortify | 2020 | Corsight, Israel, US | Israel's military tech division Unit 8200 uses this facial recognition technology, supplemented by Google Photos, to identify and detain people in Gaza. It runs in parallel to Israel's Blue Wolf program in the West Bank. | Facial recognition | US Immigration and Customs Enforcement (ICE) use Mobile Fortify app to indentify people using the nationwide IDENT biometric database (begun 1994, with data from customs, TSA, law enforcement, and elsewhere). Corsight lists other use cases as including airports, banking, construction, gaming, critical infrastructure, healthcare, prisons, smart cities, and transportation. | Corporate, Government, Press, NGO |
| Fire Weaver | 2020 | Israel, Rafael | Sensor-to-Shooter System. "The system connects all battlefield elements in real time and instantly selects the most relevant shooter for each target – enabling comprehensive situational awareness and simultaneous, precision strikes. Collecting, filtering, and disseminating data from multiple sources [...] – with details down to a specific window in a targeted building. Marking and sharing targets, enemy locations, blue forces, and POIs on all weapon sights" | Analyzing, filtering, real-time data | As-yet unknown | Corporate, Press |
| Fire Factory | 2020 | Israel, Rafael | Calculates necessary bomb strength to destroy "target" and schedules aircraft to deliver it. "uses data about military-approved targets to calculate munition loads, prioritize and assign thousands of targets to aircraft and drones, and propose a schedule." | Sorting, tasking, assign | As-yet unknown | Corporate, Press |
| Vector | 2020 | DIANA (NATO), Quantum Systems | Dual-use AI drone system (UAS) | Military surveillance, artillery fire, border patrol, mapping, and swarming | "Public safety applications include search and rescue, disaster management, live situational awareness, law enforcement, disaster control, and area surveillance." | Corporate, Government |
| Dialect Recognition Software (DIAS) | 2019 | Germany | Determines "authenticity" of migrants' claims to refugee status via computer voice analysis, used by the German Federal Asylum Agency. Claims to detect accent ("language biometry") despite high error rate. | Voice recognition, telephone audio datasets | Voice print for identification, verification, entry, and residency permit | Academic, NGO, Press |
| Automated Facial Recognition System (AFRS) | 2019 | India | Facial recognition used for missing children | Facial recognition | Repurposed for surveillance and crackdowns of peaceful protests | Press |
| Starshield, Starlink | 2024, 2019 | USA, Ukraine, SpaceX | Swarming spy satellites, imaging, reconnaisance | Infrastructure | Commercially available internet access | Press |
| Pantsir-SM | 2019 | Russia | Modifications of existing air defense systems incorporate algorithms to orient, detect, and categorize targets by degree of danger, select a method for defense, "then open fire without human intervention." | Classify, identify | As-yet unknown | Academic |
| Red Wolf | 2018 | Israel | Monitors and controls people's movement using facial recognition, scanning all faces that pass through military checkpoints | Facial recognition, camera infrastructure | Dating apps, photo library search, local policing, traffic tickets (BBC), public event security | NGO, Press |
| Blue Wolf | 2018 | Israel | Lets soldiers instantly access and add to face recognition human database (Wolf Pack) via app | App interface, facial recognition | Border patrol, security, restriction of movement | NGO, Press |
| Wolf Pack | 2018 | Israel | "Vast database containing all available information on Palestinians from the Occupied Palestinian Territories, including places of residence, family members and whether they are wanted for questioning by the Israeli authorities." | Database, data aggregation | Social credit systems, biased allocation of resources | NGO, Press |
| Mabat 2000 | 2018 | Israel, TKH Security (Dutch), Hikvision (China) | Infrastructure of surveillance cameras in occupied territories and illegally annexed areas that support the Wolf systems ~2 cameras every 5 meters | Cameras | Existing infrastructure to build AI systems on top of | NGO, Press |
| Urban Reconnaissance through Supervised Autonomy (URSA) | 2018 | USA | "enabled robots and drones to act as forward observers for platoons in urban operations. After input from the project's advisory group on ethical and legal issues, it was decided that the software would only ever designate people as "persons of interest." "DARPA program to enable improved techniques for rapidly discriminating hostile intent and filtering out threats in complex urban environments" "combining new knowledge about human behaviors, autonomy algorithms, integrated sensors, multiple sensor modalities, and measurable human responses to discriminate the subtle differences between hostile and innocent people" | Sorting, identifying, data analysis, behavior analysis | Suggested as drones to patrol cities | Government |
| Sentry | 2017 | Anduril, USA, Israel | "Sentry Tower provides clear, long range surveillance of ground-based objects of interest day or night." "Virtual border walls" | Sensors, radar,Lidar, satellite signals, edge computing, AI analysis, infrastructure, | US Customs and Border Patrol | Corporate, Press |
| TacNet | 2017 | Germany, Rheinmetall | "command and control application" including digital force deployment using AI to decide where and when to place troops. "By analyzing geo data and the monitored situation on the battlefield, the system automatically provides positions to achieve the mission objective and to eliminate potential risks. Command selects according the strategic approach and hand over information directly to his team. [...] Accelerate decision making process through AI. Share positions in real time to achieve your mission objectives as quickly as possible" | Geolocation, object detection | Tasking and tracking gig workers | Corporate |
| Bylina | 2017 | Russia, Syria | Reportedly tested in Ukraine and Syria, this system can independently evaluate the battlefield situation and calculate how to attack structures. | Info synthesis, predict, assign | As-yet unknown | Academic |
| Maven Smart System | 2017 | USA, Google, Microsoft, Verizon, Facebook, Skype, YouTube, Apple, Yahoo, etc., later Palantir | The results of the widely publicized PROJECT MAVEN, this platform automates target identification from satellite images, using AI image classification. Object detection from satellite images to destroy ground targets. "a Pentagon program that developed target recognition algorithms for video footage from drones" "Through Maven Smart System, software and advanced digital capabilities act as the connective tissue between operators and sensor feeds, platforms (hardware and software), and algorithms. These AI-enabled capabilities drastically enhance the speed and accuracy of software-enabled workflows, improving operational decision-making through human-machine teaming." | Object detection, computer vision, data processing | Mapping, location confirmation, surveillance | Press |
| PackBot | 2016 | Arlington Capital Partners (iRobot) | Search and rescue, and bomb disposal robot | Site mapping, sensors | Roomba | Press |
| GIS Arta | 2014 | Ukraine | Pairs a battlefield target to an artillery unit ("like Uber rider to driver" | Tasking, sorting | Driver and delivery apps, logistics apps | Press |
| IrisGuard | 2013 | Jordan, UNHCR, IrisGuard Inc | Iris scanning biometric hardware, software, and storage | Iris scanning biometric hardware, software, and storage | "Blue-Collar Salary Assurance, Mobile Wallets, SIM Card Registration, Digital ID, Elections, Work Visa Validation, Notary Public, Document Delivery, Medical Education Exams, Electronic Medical Record, Patient Reunite Service, Clinical Research, Pension & Social Welfare, Cash, Food and NFI Assistance, Card Validation, Border Control & eGates" | Academic, Press |
| Skynet (US) | 2012 | USA | "uses phone location and call metadata from bulk phone call records to detect suspicious patterns in the physical movements of suspects and their communication habits, according to a 2012 government presentation The Intercept obtained from Edward Snowden. The presentation indicates that Skynet looks for terrorist connections based on questions such as “who has traveled from Peshawar to Faisalabad or Lahore (and back) in the past month? Who does the traveler call when he arrives?” It also looks for suspicious behaviors such as someone who engages in “excessive SIM or handset swapping" or receives “incoming calls only." The goal is to identify people who move around in a pattern similar to Al Qaeda couriers who are used to pass communication and intelligence between the group's senior leaders. " | Filtering, sorting, pattern matching | Leads to AI-DSS used above ^^ | Press |
About
Winner of the 2025 STARTS Grand Prize – Artistic Exploration
