# Project Maven and Kill Chain Compression ## BLUF Project Maven (formally the Algorithmic Warfare Cross-Functional Team, AWCFT) is the US Department of Defense initiative that first operationalised AI at scale for military targeting — and its successor, the **Maven Smart System (MSS)**, is now the primary AI backbone of US joint targeting and intelligence operations across active theatres. Launched in 2017 as a response to the proliferation of drone footage vastly exceeding human analytical capacity, Maven established the DoD's foundational doctrine for human-machine teaming in the kill chain. The MSS — operated by [[Palantir Technologies]] — represents the mature operational deployment of Maven's original concept, and its integration into JADC2 architecture has made AI-assisted targeting a structural feature of US military operations rather than an experimental programme. --- ## Origins and Strategic Context Project Maven was established by Deputy Secretary of Defense [[Robert Work]] in April 2017. The driving operational problem: US and coalition forces in Iraq and Syria were generating approximately 1.8 million minutes of full-motion video from ISR platforms every day. Human analysts could not process a fraction of this feed in tactically relevant timeframes. Maven's initial mandate was narrow: apply computer vision algorithms to identify objects of interest (vehicles, structures, human activity patterns) in FMV automatically, cueing human analysts to review flagged segments rather than entire feeds. The concept was deceptively modest. Its implications were not. **The strategic insight:** If AI could compress the *analysis* phase of the kill chain, the sensor-to-shooter timeline could be reduced from hours to minutes. At sufficient scale, this would create a qualitatively different kind of military — one that could process the entire electromagnetic and visual environment of a theatre in near-real-time. --- ## The Google Controversy and Its Strategic Significance In 2018, approximately 3,000 Google employees signed a letter objecting to the company's Maven contract, demanding Google "not be in the business of war." The subsequent Google withdrawal from Maven renewal negotiations was publicly framed as an ethical stand. Its strategic significance was different: it demonstrated that **Silicon Valley's social consensus** could constrain DoD access to frontier AI capability — a structural vulnerability the Pentagon subsequently worked systematically to address through the creation of the Joint AI Center (JAIC), the Chief Digital and AI Office (CDAO), and the cultivation of Palantir, Anduril, and Scale AI as defence-native AI vendors. The controversy also revealed the fundamental tension that has never been resolved: the most capable AI development happens in the commercial sector; that sector's workforce has political opinions about weapons applications; and the DoD has no domestic state AI capability equivalent to what the PLA can compel from Chinese commercial entities. --- ## Maven Smart System (MSS) — Mature Deployment The Maven Smart System, operated by Palantir Technologies since 2020, is the operational evolution of the original Project Maven concept into a full-spectrum military AI platform: | Function | Capability | |---|---| | Multi-domain intelligence fusion | Synthesises SIGINT, IMINT, HUMINT, OSINT into unified operational picture | | Automated object/activity recognition | Computer vision at theatre scale across all ISR feeds simultaneously | | Targeting matrix generation | Automated prioritisation and weapon-to-target matching recommendations | | Pattern of life analysis | Longitudinal tracking of target behaviour, movement, and network associations | | Battle damage assessment | Post-strike automated analysis of imagery to assess effect | | JADC2 integration | Feeds directly into Joint All-Domain Command and Control decision architecture | MSS is deployed in the CENTCOM AOR (Iraq, Syria, broader Middle East), was used in support of operations in Ukraine (intelligence sharing), and is assessed as operational in Pacific Command planning. --- ## Kill Chain Compression: The Doctrinal Shift The traditional OODA Loop (Observe → Orient → Decide → Act) in military operations assumed human cognitive processing at each stage. The introduction of AI into the observe and orient phases fundamentally disrupts this assumption: **Traditional kill chain:** Sensor collection → Human analysis (hours) → Targeting recommendation → Human review → Strike authorisation → Execution **Maven/MSS kill chain:** Sensor collection → AI analysis (seconds/minutes) → Automated targeting matrix → Human approval → Execution The human role has shifted from **analyst** to **approver**. The practical implication — visible in both the IDF's Lavender system and the MSS deployment — is that the volume of decisions processed and the speed at which they are processed exceed human deliberative capacity. Meaningful human review per targeting decision is structurally compromised by the throughput the system enables. --- ## Strategic Implications **For US Military:** The MSS/Maven architecture creates capability dependencies that cannot be walked back. The DoD's targeting infrastructure is now AI-native — not AI-assisted. Removing the AI layer would collapse operational tempo to pre-2017 standards. **For Adversaries:** The primary vulnerability is not hacking Maven's servers. It is corrupting the intelligence feeds entering it — data poisoning of SIGINT inputs, false pattern-of-life data, fabricated imagery metadata. The AI processes what it is given; adversarial manipulation of inputs achieves targeting degradation without ever touching the system itself. **For Allies and Partners:** MSS integration creates interoperability dependencies with US targeting architecture that constrain allied autonomy in joint operations. Nations operating inside a US-led JADC2 framework are, in practice, operating under Palantir's epistemological layer. --- ## Key Connections - [[01 Actors & Entities/14_Corporations_&_Tech/Palantir Technologies]] - [[01 Actors & Entities/16_Leaders_&_Figures/Robert Work]] - [[02 Concepts & Tactics/Algorithmic Warfare]] - [[02 Concepts & Tactics/Kill Chain]] - [[02 Concepts & Tactics/Intelligentised Warfare]] - [[01 Actors & Entities/13_Agencies_&_Departments/Israel Defense Forces]] - [[Publish/Quartz/Palantir — The Company That Owns the Western Kill Chain]] --- ## Sources 1. DoD — Project Maven establishment memo, Deputy SecDef Work (April 2017) 2. Google — Internal employee petition and subsequent Maven contract non-renewal (2018) 3. DoD CDAO — Maven Smart System public disclosures (2020–2024) 4. Defense One — "AI Is Already at War" series (2023–2024) 5. CNAS — "AI and the Future of Defense" (2023)