A look at Project Maven, the US DOD's flagship AI effort which identifies battlefield targets, and at concerns, including adversaries poisoning training data
The near-term relevance is clearer after US Central Command said Maven-developed machine learning helped narrow targets for more than 85 air strikes. That makes data integrity and human review consequential operational issues, not merely model-development concerns.
First-order effects
The DOD and military units using Project Maven must treat training-data poisoning as a direct risk to target-identification outputs, increasing the importance of validating data and reviewing recommendations.
Personnel responsible for battlefield targeting face greater pressure to establish when an AI-generated lead is reliable enough to inform a strike decision.
Second-order effects
Defense AI suppliers and prospective Pentagon partners will be judged not only on model capability but also on data provenance, resilience testing, and auditability in deployed systems.
Operational adoption can shift spending and attention toward secure data pipelines and verification processes, rather than treating the model itself as the sole product.
Third-order effects
If AI-assisted targeting becomes routine, military AI procurement is likely to center increasingly on accountable deployment and adversarial robustness, with data control becoming a strategic capability.
The pattern points to a durable tension in dual-use AI: faster decision support can raise the cost of errors or manipulation, making human oversight and system assurance central to legitimacy.
The trend: Project Maven is one instance of defense AI moving from experimentation to operational decision support, where trust in data and deployment controls matters as much as model performance.
Even the most skeptical of US military operators are embracing artificial intelligence to identify targets on the battlefield. If they keep going, war could speed up faster than humans can think Read The Big Take ⬇️ https://www.bloomberg.com/...
AI identifies targets and supplies Maven Smart System, along with other data feeds. A “tactical data link” can transmit a commander's decision to fire between MSS and a weapon.
The Army source quoted in this article says Project Maven, currently being used to help blow things up in Yemen, Iraq, and Ukraine, has a 60% accuracy rate w/ identifying objects, as sometimes “the system confuses a truck with a tree or ravine” https://www.bloomberg.com/...
The US military has embraced AI to locate targets they've gone on to strike — a confirmation that demonstrates AI warfare is already here. @KatrinaManson explains https://www.bloomberg.com/... [video]
As a result of Project Maven, the Pentagon's controversial 2017 effort to bring AI to combat, I learned the US is already using AI targeting to find enemy targets they have then struck with weapons fire, in Iraq and Syria.
The US military's Maven system for identifying battlefield targets using AI is only accurate 30-60% of the time, depending on conditions. https://www.bloomberg.com/... [image]
.@KatrinaManson just published a feature on how the US military is using AI to identify strike targets. Never seen an account this detailed. It's a must-read. Gift link here: https://www.bloomberg.com/...
Trusting the computer entirely would be faster still, but a senior targeting officer said that would introduce errors. “I don't ever wanna get caught short. We get caught short, we're screwed.” https://www.bloomberg.com/...
I remember distinctly in 2019 when I was dismissed by a US think tanker for suggesting that Project Maven would be used for precisely this kind of accelerated targeting. This speed with which goal post change should concern us all (1/4) https://www.bloomberg.com/...
AI warfare deep dive I finally got to see Maven Smart System, the US military platform that ingests Maven AI targeting data and other feeds, and which I learned operates in more than 100 locations, from the Middle East to the White House. https://www.bloomberg.com/...