Home / The AI Kill Chain
F2T2EA Framework

The AI Kill Chain

The kill chain describes every step between detecting a target and destroying it. AI is compressing that process from 24 hours to 3 seconds — and removing humans from the loop at each phase.

FIND — FIX — TRACK — TARGET — ENGAGE — ASSESS

The F2T2EA Framework

Six Phases from Sensor to Strike

Originally a US military targeting framework, F2T2EA defines the full sequence of actions required to engage a target. Each phase has been dramatically transformed — or is being transformed — by artificial intelligence and autonomous systems.

Phase 01
Find
Sensor Detection
Traditional
4–8 hours — human analysts reviewing satellite imagery and signals intelligence
AI-Assisted
15 minutes — Palantir Maven auto-detects objects across terabytes of ISR footage
Fully Autonomous
0.3 seconds — onboard computer vision classifies targets in real-time
Key Systems
Satellite ISR Drone Surveillance SIGINT Collection Social Media OSINT Maven Smart System
Phase 02
Fix
Precise Location
Traditional
2–4 hours — coordinate confirmation from multiple intelligence reports
AI-Assisted
5 minutes — multi-source fusion correlates signals across sensors
Fully Autonomous
0.5 seconds — GPS + optical visual lock, onboard geolocation fused
Key Systems
Geolocation AI Pattern-of-Life Analysis Multi-Source Fusion GPS + Visual Lock
Phase 03
Track
Continuous Monitoring
Traditional
Continuous human monitoring — requires dedicated analyst teams, subject to fatigue and attention gaps
AI-Assisted
Automated tracking with exception-based alerts — analyst engaged only on movement events
Fully Autonomous
Predictive movement modeling — AI anticipates target trajectory without human input
Key Systems
Persistent Surveillance Drones Low-Earth Orbit Constellations Predictive Analytics
Phase 04
Target
Decision to Engage
Traditional
Hours to days — commander review, legal review, Rules of Engagement verification, escalation chain
AI-Assisted
Minutes — AI generates targeting recommendation, human commander approves
Fully Autonomous
Milliseconds — algorithm decides to engage based on pre-programmed criteria

Critical Decision Point

This is where the human-in-the-loop debate lives. Who — or what — makes the kill decision? This single phase determines whether a weapon system complies with International Humanitarian Law.
Key Actors
Human Commander AI Targeting System Legal Review (JAG) Autonomous Algorithm
Phase 05
Engage
Weapons Release
Traditional
Seconds to minutes — pilot or ground operator manually fires; dependent on platform availability
AI-Assisted
AI-guided terminal phase — human fires but AI guides munition autonomously to target
Fully Autonomous
Self-directed engagement — loitering munition selects and strikes without further human input
Key Systems
Loitering Munitions AI-Guided Missiles Autonomous Strike Drones Lancet / Switchblade
Phase 06
Assess
Battle Damage Assessment
Traditional
Hours — post-strike reconnaissance sortie required; imagery reviewed by human analysts
AI-Assisted
Seconds — real-time video feed with AI damage classification and re-engagement recommendation
Fully Autonomous
Instant — onboard sensors assess strike outcome; system automatically loops back to FIND for re-engagement
Key Systems
EO/IR Sensor Suite AI Video Analysis Automated Reporting Re-engagement Loop
Speed Comparison

Full Chain Time: Sensor to Strike

Across all six phases combined, AI doesn't just accelerate decision-making — it fundamentally changes the nature of human control. When the entire chain runs in under 3 seconds, human oversight becomes operationally impossible.

Traditional
Full Human Chain
24+ Hours
AI-Assisted
Human-on-the-Loop
~45 Minutes
Fully Autonomous
No Human Control
Under 3 Seconds
Real-World Examples

Kill Chain Case Studies

How the F2T2EA framework has been applied in documented or publicly discussed operations, and what they reveal about the trajectory toward full autonomy.

IDF — Gaza Conflict 2023-2024
Israel Lavender System
FIND
AI mass-classification of intelligence dossiers
~20 sec
FIX
AI correlates identifier to geolocation data
~20 sec
TRACK
Automated pattern-of-life monitoring
Continuous
TARGET
Human officer approves AI recommendation — reported 20-second review
~20 sec
ENGAGE
Human-authorized strike with AI-guided munitions
Minutes
ASSESS
AI-assisted post-strike review
Seconds
Reported by +972 Magazine and confirmed by current/former IDF intelligence officers. Lavender generated targeting lists at mass scale. The nominal human-in-the-loop was described as rubber-stamp level, raising fundamental questions about meaningful human control under IHL.
USAF / CENTCOM Operations
MQ-9 Reaper Chain
FIND
Human analysts review drone and satellite ISR feeds
2–6 hrs
FIX
Human intelligence correlation and coordinate confirmation
1–3 hrs
TRACK
Continuous human-piloted persistent surveillance
Hours+
TARGET
Full JAG review, commander authorization, LOAC compliance check
Hours–Days
ENGAGE
Human pilot fires; AI-assisted terminal guidance on Hellfire/GBU
Seconds
ASSESS
Human-reviewed post-strike ISR, AI-assisted BDA classification
Hours
Current standard US military practice. Human control retained throughout most phases, but AI-assisted terminal guidance already deployed. This is the benchmark that autonomous systems are being compared against — and rapidly replacing.
Projected Capability — 2027+
Full Autonomous Chain
FIND
Onboard CV classifies target from sensor array
0.3 sec
FIX
GPS + visual lock, multi-source fusion
0.5 sec
TRACK
Predictive movement model, no human monitoring
Continuous
TARGET
Algorithm applies pre-programmed ROE parameters — no human
<1 ms
ENGAGE
Self-directed strike, no operator
<1 sec
ASSESS
Sensor-based auto-assessment; re-engage or stand down
Instant
No currently deployed system publicly claims this full chain. However, Lancet-4 (Russia), Loyal Wingman CCA concepts (US), and next-generation loitering munitions are progressively approaching it. The legal and ethical framework to govern this capability does not yet exist.
Related Analysis
Deep Dive
AI Ethics & IHL
Read
Analysis
Military Doctrine
Read
Risk Assessment
WW3 Probability
Read
Threat Matrix
Autonomous Threats
Read