IAF Takes Cautious Approach to AI in Combat Aviation, Prioritizes Human-AI Collaboration

The Indian Air Force (IAF) is cautiously exploring the potential of artificial intelligence (AI) in combat aviation, with a focus on developing unmanned "loyal wingmen" to complement manned fighter jets. While AI-controlled fighter jets like the US Air Force's F-16 "Vista" have shown promise, the IAF is taking a more measured approach, prioritizing collaboration between humans and AI.

The IAF is particularly interested in the concept of unmanned aerial vehicles (UAVs) working in tandem with manned fighter jets, known as "loyal wingmen." India's indigenous development of the CATS Warrior Loyal Wingman, designed by state-owned HAL, exemplifies this collaborative approach. These unmanned aircraft can perform various tasks, such as reconnaissance, electronic warfare, and even combat roles, while being controlled by a human pilot in a nearby fighter jet.

While the IAF acknowledges the potential of AI in air combat, it is not yet ready to fully embrace autonomous fighter jets. The focus is on incorporating advanced automation features into manned aircraft, such as auto take-off and landing systems, to reduce pilot workload and enhance operational efficiency. India's ambitious AMCA (Advanced Medium Combat Aircraft) program is a prime example of this approach, prioritizing automation over full autonomy.

The IAF's cautious approach to AI stems from concerns about the technology's current limitations and potential unpredictability. AI algorithms may struggle to handle complex and dynamic combat scenarios, and there are ethical considerations regarding the use of lethal autonomous weapons systems. Additionally, the IAF leadership may be more comfortable with a pilot-centric approach, gradually incorporating AI as a supporting tool.

As AI technology continues to advance, the IAF's approach may evolve. A generational shift within the IAF ranks could potentially lead to a more open embrace of AI-controlled fighter jets in the future. However, for now, the focus remains on developing unmanned loyal wingmen and incorporating advanced automation into manned aircraft.

 

[Ranjeet Singh]

Why not we ink deal with USA to make F-21+ F-36+ F-21 Vista here in India ? We must ink this deal ! F-21 with laser weapon Pod, Super cruise, triple and double missile launch adapters and F-35 type AESA RADAR and Ew suite and later on F-36 should be made in India !

 

[Bhishma]

We need to master the AI tech. It will hit the battlefield very soon. AI is growing by leaps and bounds. Compared to the AI problems getting solved these days, autonomous flight & fighter maneuvers are relatively simple. 5 years from now, that tech will be very mature. We need to start now in order to be there. We just have to master this technology using cheap supersonic drones. Then we can fit it into our old fighters like MiG21, worst is losing the jet, which we are going to phase out anyway. This tech will be very important for Ghatak/CATS warrior. Its doable. We need to start now.

 

[Sandeep]

It has taken the USA over 50 years to look at aerial warfare, tactics, predictions, enemy jet studies and weapons capability before they could develop a AI driven pilot. To develop that it's very complex and complicated as you have to get the software, hardware and flight laws absolutely perfect. India has not reached that advanced stage yet. We need to get our manned jets right with the Tejas MK1A and MK2 and AMCA and then try pairing it with the CATS program. After studying the relationship and capabilities between the two then we can think about deploying a unmanned jet/drone.

 

[Bhishma]

It has taken the USA over 50 years to look at aerial warfare, tactics, predictions, enemy jet studies and weapons capability before they could develop a AI driven pilot. To develop that it's very complex and complicated as you have to get the software, hardware and flight laws absolutely perfect. India has not reached that advanced stage yet. We need to get our manned jets right with the Tejas MK1A and MK2 and AMCA and then try pairing it with the CATS program. After studying the relationship and capabilities between the two then we can think about deploying a unmanned jet/drone.

It doesnt matter. We can do it. Anybody can do it. Plenty of open source code out there. AI community publishes lot of papers. What you said is not relevant. Two minute papers posted a video recently (titled "DeepMind’s New Robots: An AI Revolution!") - they showed how AI players learned to play football in virtual environment. Then they were able to transfer that learning to robots & play football in real world. The problem we are dealing with has about the same complexity. Its doable. China will be 100% doing it. They are making big investments in AI. Are we going to get left behind ? AI is growing very rapidly. Only a matter of time (3-5y) before China puts them in swarm drones and uses it in offensive ops.

 

[Sandeep]

It doesnt matter. We can do it. Anybody can do it. Plenty of open source code out there. AI community publishes lot of papers. What you said is not relevant. Two minute papers posted a video recently (titled "DeepMind’s New Robots: An AI Revolution!") - they showed how AI players learned to play football in virtual environment. Then they were able to transfer that learning to robots & play football in real world. The problem we are dealing with has about the same complexity. Its doable. China will be 100% doing it. They are making big investments in AI. Are we going to get left behind ? AI is growing very rapidly. Only a matter of time (3-5y) before China puts them in swarm drones and uses it in offensive ops.

The problem with using open source code leaves your system vulnerable to hacks, faults and problems. It has to be a completely indigenous software as the platform/jet won't be the same for everyone.