US Air Force Explores AI to Spot Errors in Mid‑Air Refuelling

The United States Air Force (USAF) is investigating the use of artificial intelligence (AI) to detect mistakes during aerial refuelling missions and to improve the feedback given to crews.

How the system would work

• Video capture – Cameras on KC‑46 Pegasus tankers would record the entire refuelling operation.
• Computer vision – AI would analyse the footage in real time, tracking aircraft speed, alignment, separation and boom movement.
• Performance reports – The software would generate graphical overlays, colour‑coded maps and a chronological log of key events, allowing instructors to pinpoint specific errors.

Current assessment challenges

Presently, evaluations rely heavily on instructor observations, crew self‑reports and manual review of recorded footage. These methods can lead to inconsistent judgments, especially when brief errors are missed or interpreted differently by different observers.

Expected benefits

• Consistent metrics – Automated measurements reduce human bias.
• Detailed feedback – Pilots and instructors can see exactly where a manoeuvre deviated from the ideal.
• Training integration – The system could be linked to flight simulators, virtual reality (VR) and augmented reality (AR) platforms already used in aviation training.

Development timeline

The project is part of the Small Business Innovation Research (SBIR) programme. Companies interested in developing the technology have until 25 November 2026 to submit proposals.

Scope beyond refuelling

While the initial focus is on the KC‑46 Pegasus, the solicitation also invites solutions that could be applied to other aviation activities within the Department of Defense.

How to verify the announcement

The information comes from a USAF research solicitation released on the Department of Defense’s official website. Interested developers can review the full solicitation document and submit proposals through the SBIR portal.

Practical implications

If successful, the AI system could:

• Reduce the time required for post‑mission debriefs.
• Lower the risk of human error during one of aviation’s most demanding procedures.
• Provide a scalable model for automated performance assessment across the military.

The initiative reflects the USAF’s broader push to incorporate advanced analytics and machine learning into training and operational processes.