Bot shots: US Army enlists AI startup to provide target-tracking
Because handing battlefield ID to an algorithm has never gone wrong before, right?
The US Army is preparing to deploy a new AI product that promises to automatically identify and track potential targets on the battlefield. However, humans will continue to make life and death decisions.
San Francisco AI startup TurbineOne announced on Thursday that the Army has awarded it a five-year contract worth up to $98.9 million to take its flagship Frontline Perception System (FPS) from small-scale pilot projects to what a company spokesperson described to The Register as a "production-scale" contract.
TurbineOne describes FPS as a model-agnostic machine learning platform that automates the assessment of military targets. The platform ingests data from multiple sensors, such as air, land, sea, and space-based imagery and signals, to detect battlefield threats like drones, enemy positions, or other targets.
FPS does all of that in a no-code, hardware-agnostic environment that lets the average soldier in the field "build, retrain, and deploy custom machine learning models at the edge without coding," according to the company. Most critically, FPS is designed to operate without a connection to the internet or cloud services. Everything runs on local devices, and soldiers can even get actionable intelligence delivered to smartphones and other mobile devices.
US government contracting data suggests TurbineOne has been testing FPS with various military branches since it opened its doors in 2021, but not all of its applications have focused purely on the battlefield. It's also been designed to do things like spot debris and other foreign objects on military airfields to prevent accidents.
- Defense Department signs OpenAI for $200 million 'frontier AI' pilot project
- US military pulls the trigger, uses AI to target air strikes
- Defense AI models 'a risk to life' alleges spurned tech firm
- US Army turns to 'Scylla' AI to protect depot
"Despite advancements, current military threat identification processes remain extremely resource-intensive, costly, and difficult to scale," TurbineOne wrote in a blog post in May announcing a $36 million Series B funding round. AI, of course, is the solution.
"By providing a no-code, easy-to-use toolkit, TurbineOne ensures that warfighters can rapidly deploy powerful AI tools themselves, without waiting for specialized technical support," TurbineOne said. The company told us that FPS is already in the hands of Army units in multiple theaters, and it hopes the deal announced on Thursday will speed its deployment across the military.
Bringing AI onto a battlefield and trusting it to identify targets - something humans often struggle with in combat situations - raises lots of questions about safety and responsibility, naturally, but TurbineOne had a ready answer when asked about those particular risks.
According to a company spokesperson, FPS doesn't so much suggest who to kill as it assists in searching for, identifying and tracking targets using data from multiple sensors available to its operators, who then make the final decision about whether a target is a threat or not. The company added that field retrainability can help FPS adapt to changing situations without soldiers having to assume it's always correct, which TurbineOne said conforms to the DoD's responsible AI principles.
TurbineOne also told us that it has one of the fastest records for a small business, going from the pilot phase to a full-scale DoD contract, which it said proves FPS' capabilities. One could also see that rapid adoption timeline as a sign the Pentagon is falling over itself to adopt private-sector technology that's still in its infancy and is facing a growing list of unforeseen consequences AI firms have yet to address. ®