DARPA and the Air Force Fly an F-16 With No One at the Controls

DARPA and the US Air Force flew a specially modified F-16 under the VENOM (Viper Experimentation and Next-generation Operations Model) program with an AI agent controlling flight for extended stretches — no human input on the stick, with a test pilot aboard solely to monitor and ensure mission objectives were met.

How the Modification Works

The VENOM jets use an aftermarket kit letting a pilot switch between traditional human controls and autonomous AI flight with the flip of a switch, rather than requiring a purpose-built uncrewed airframe. This first autonomous flight took place at Eglin Air Force Base in Florida, with DARPA describing it as a genuinely real-world test, not a simulation.

Why This Is Relevant Outside Defense

• The test represents one of the clearest public demonstrations of an AI agent making continuous, high-stakes, real-time control decisions with a human present only to intervene, not to act
• DARPA has said upcoming test campaigns will push the AI pilots into progressively more realistic flight conditions, expanding the operational envelope over time rather than stopping at this milestone
• For any industry evaluating how much autonomous decision authority to hand an AI system, the human-monitor-not-operator model DARPA is using here is a concrete design pattern worth studying, independent of the military application