Command authority will remain paramount: Lt Gen Shinghal at AI Summit 2026

Speaking at the India AI Impact Summit, he said that it is important for the military to decide which decisions must remain with commanders and which can be delegated to AI

In an era where artificial intelligence (AI) is being widely used across sectors, Lt Gen Vipun Singhal said that no matter how accurate and proficient AI becomes, human judgment, moral responsibility and command authority must remain paramount in military matters.

 

Speaking at the India AI Impact Summit 2026, Lt Gen Singhal said that while AI can assist in decision-making, only humans can exercise judgment. “AI can inform decisions, but only humans can exercise judgment and bear responsibility for them. AI today has unprecedented speed, scale and efficiency. Those attributes are real. What deserves deeper reflection is something else — the new burdens AI places on leadership, command responsibility and strategic stability,” he said.

 

 

He outlined four questions that require deeper reflection within the armed forces. He said it is essential to determine which decisions must remain with commanders and which can be delegated to algorithms. “Firstly, which decisions must remain with the commanders, and which decisions can be delegated to data or algorithms? No matter how accurate or proficient they become, human judgment, moral responsibility and command authority remain paramount,” he said.

 

Lt Gen Singhal also raised concerns about the training of AI models and the quality of data they rely on. “Have our rules of engagement and safety frameworks kept pace with systems that rely on data? Do we fully understand how these systems are trained, the quality of their data and the assumptions that they make?” he asked.

 

Trial of AI systems

 

While conventional military weapon systems undergo extensive trials and evaluation before deployment, Lt Gen Singhal questioned whether AI-enabled systems are subjected to similar scrutiny. “Are we subjecting AI-enabled systems to the same rigour that we subject new weapon systems to — trialling them, evaluating them, fielding them in contested conditions, and then applying them? After all, an AI-enabled system is not software; it is a weapon,” he said.

 

He recounted an incident in which human judgment overrode an AI-generated recommendation during a high-tempo military operation. “During a recent high-tempo military operation, a senior commander was presented with a machine-generated recommendation based on multiple sensor threats and AI analysis to engage a target immediately. The system was confident. The probability score was high. The decision window was measured in seconds. But the commander paused,” he said.

 

The commander questioned what the system might not have accounted for. The area under surveillance had adversary troops, but a civilian evacuation had just begun and had not yet been incorporated into the system’s data. Exercising restraint, the commander halted the strike, preventing potential civilian casualties. The mission was later completed successfully.

 

Lt Gen Singhal also referred to India’s AI governance guidelines, describing them as a pragmatic framework that recognises AI’s probabilistic, generative and adaptive nature, which may lead to unintended consequences.

 

“I must say, military leaders must stay on the learning curve, understand technology — both its challenges, pitfalls and uses — while also ensuring that technology is absorbed and adopted by our units and groups,” he said.

 

On AI adoption in the Indian armed forces, he said, “Let me be very clear that we as the Indian armed forces and the Indian Army are fully cognisant of the transformative power of AI to increase our operational efficiency, and we are making every effort in that direction to ensure that AI is incorporated into our decision support systems, into our surveillance, reconnaissance and all the other functions that we perform.”

 

Emphasising the enduring role of human intelligence, he said that history has shown human wisdom can respond effectively to new technological challenges. “The rules governing the use of nuclear, biological and chemical weapons; the framework that has come into place for governing landmines; the Geneva Conventions themselves — which talk about conduct — have all stood the test of time, even in conflict. So I am confident that AI, as the new looming technology on the forefront, will also find ways and means to be regulated,” he said.