AI at the controls: US Air Force partners with MIT to train the next generation of test pilots
September 20, 2025
Artificial intelligence (AI) is no longer a distant prospect for the US Air Force; it is swiftly becoming a fundamental force multiplier in global air and space operations.
But as AI systems grow more capable, the challenge shifts from merely developing these technologies to ensuring that Airmen and test pilots are equipped to evaluate, integrate, and lead AI-enabled operations with the highest standards of safety and effectiveness.
In August 2025, the US Air Force Test Pilot School (TPS), in collaboration with the Department of the Air Force–Massachusetts Institute of Technology Artificial Intelligence Accelerator (DAF-MIT AIA), conducted an intensive two-week workshop designed to do exactly that.
Hosted at MIT’s cutting-edge Computer Science and Artificial Intelligence Laboratory (CSAIL) and Lincoln Laboratory’s Beaver Works, the programme brought together 40 of the Air Force’s most skilled pilots, engineers, and technicians from platforms as varied as the MQ-9 Reaper drone, the B-52 Stratofortress bomber, and the F-35 Lightning II stealth fighter.
From cockpits to code: AI training for test pilots
The modern battlefield and aerospace environment demand that test pilots and engineers understand not just aerodynamics and avionics, but increasingly complex AI systems that will play a central role in future missions.
Unlike traditional aircraft training, this workshop focused on equipping these operators with the knowledge to design, test, and critically assess AI-enabled technologies, such as autonomy algorithms and machine learning tools for anomaly detection.
Maj. Morgan Mitchell, the workshop manager for DAF-MIT AIA, highlighted the imperative of this new kind of training:
“Our testers need to lead the integration of AI and machine learning into Air Force operations. Understanding how these systems function in real-world scenarios and their limitations is fundamental to our mission success.”
During the course, participants engaged hands-on with MIT’s RACECAR platform, an AI-powered autonomous vehicle testbed, to experiment with autonomy concepts in a controlled but realistic environment. They also explored MIT’s CSAIL Robot Apartment Living Lab, which simulates complex, dynamic spaces to test AI navigation and decision-making in real time.
Human-machine teaming in Air Force operations
Central to the Air Force’s AI strategy is the concept of Human-Machine Teaming (HMT), where AI systems augment, rather than replace, human decision-making.
AI excels at rapidly processing vast amounts of data, identifying patterns, and suggesting actions. However, it lacks the intuition, contextual awareness, and ethical reasoning unique to trained Airmen.
This delicate balance means that AI acts as a force multiplier, amplifying human capabilities while human judgment remains paramount, particularly in high-stakes or strategically sensitive situations.
For example, AI can rapidly sift through sensor data to detect potential threats or anomalies, but it is ultimately an Airman who must evaluate these findings, apply contextual knowledge, and make operational decisions.
Col. Scott Ruppel, Director of the DAF-MIT AI Accelerator, explained the importance of this partnership:
“Our goal is to prepare test pilots and engineers to work side-by-side with AI-enabled systems, ensuring they can test, validate, and oversee these technologies under real operational conditions in both air and space domains.”
Real-world applications of AI in the US Air Force
The USAF is already deploying AI to great effect across various operational areas.
- Predictive maintenance: AI-enabled systems analyse aircraft sensor data to anticipate mechanical failures before they occur, significantly reducing downtime and costs.
- Crew optimisation: The C-17 Globemaster III fleet uses AI-driven scheduling tools to streamline assignments, improving efficiency.
- Autonomous systems: AI enhances the MQ-9 Reaper drone’s surveillance and strike capabilities in contested environments.
- Cybersecurity: AI algorithms detect unusual network activity patterns, enabling faster responses to potential cyber threats.
Training programmes like the one at MIT ensure that Airmen not only understand how these AI applications work but also how to push their limits safely during testing and evaluation.
Building an AI training pipeline through MIT and Stanford
The TPS workshop is one element of the broader and historic collaboration between MIT and the Department of the Air Force, formalised in 2019 with the establishment of the DAF-MIT AI Accelerator.
This partnership fosters a two-way exchange: Airmen embedded at MIT contribute operational insights to researchers, while MIT’s breakthroughs in AI research rapidly translate into practical applications for the Air Force.
The Accelerator has overseen more than 20 interdisciplinary projects spanning autonomy, machine learning, AI ethics, and human-machine interaction. Its unique model integrates officers and enlisted Airmen directly into MIT’s research environment, ensuring AI development is grounded in real-world military requirements.
Furthermore, the AI training programme at TPS will now rotate biannually between MIT in the summer and Stanford University in the winter. This guarantees that each new cohort of test pilots and engineers gains exposure to cutting-edge AI advances from multiple world-leading institutions.
Preparing for the future of AI in air and space
The Air Force recognises that AI’s role will only expand, from battlefield decision support and autonomous air vehicles to space domain operations and cyber defence.
As such, cultivating a cadre of operators who can critically assess, validate, and direct AI systems is vital to maintaining strategic advantage.
This training paradigm also underscores a critical philosophical shift: while AI offers unprecedented capabilities, human judgment remains the cornerstone of mission success.
The USAF’s investment in training test pilots and engineers at MIT reflects a commitment not just to AI innovation but to responsible, ethical deployment, ensuring that AI enhances rather than replaces the human element.
As Maj. Mitchell puts it, “This is about preparing our people to lead—not just follow—the integration of AI into every facet of Air Force operations.”
