“When you’re such a large market and so reliant on various devices, you also need to have some degree of strategic autonomy. This is a concern not just for India. Europe and the US do it. We’ve started doing it in some sectors like telecom and for CCTV cameras. We may have to expand it to a few more areas particularly in the context of AI,” Krishnan said at an event organised by policy think-tank National Council of Applied Economic Research.
AI’s increasing footprint on disparate sectors from manufacturing to agriculture, is enabled by the collection of huge amount of data through sensors and other Internet of Things (IoT) devices. These devices often end up being black boxes with no access and the government needs to ensure they come in from trusted sources, Krishnan said. This needs to be analysed from the perspective of industrial espionage to strategic concerns, he added.
Krishnan cautioned the problem extended to the digital public infrastructure (DPI) stack of government services involving biometric authentication done through a range of devices. “Both for AI and DPI, the integrity of the electronic systems and the way that runs is becoming more and more relevant.”
In April, 2024, the Ministry of Electronics and Information Technology introduced essential requirements for CCTV systems, mandating transparency on the origin of critical hardware like the system-on-chip and rigorous lab testing against unauthorised remote access. Under the directions, government departments are also restricted from buying non-certified surveillance hardware to prevent data from being leaked via foreign-manufactured firmware or hidden backdoors.
Similarly, under the national security directive, telecom operators are mandated to only source equipment from government-designated trusted sources and trusted products.