Union Home Minister Amit Shah has told India’s security establishment to lean harder on artificial intelligence — but not in the sweeping “AI will fight terrorists now” way a lot of headlines are making it sound.
Speaking at the 9th National Security Strategies Conference (NSSC) in New Delhi on August 24, Shah directed security agencies to run AI-based analysis of information sitting on the Multi-Agency Centre (MAC) platform — the Intelligence Bureau’s shared intelligence-fusion system — and to use that analysis to build Standard Operating Procedures for tackling emerging national security threats. The two-day conference, held in hybrid format, brought together more than 850 officials, including heads of Central Armed Police Forces, Deputy National Security Advisors, and state police chiefs.
If that sounds familiar, it’s because it is, in part. Shah gave a similar push for “Big Data and AI-driven analytics” at MAC back in July 2025. This week’s directive reads less like a new policy launch and more like Delhi tightening the screws on something already in motion.
What Was Actually Announced
Here’s the plain version, stripped of the “AI vs terrorism” framing you’ll see elsewhere:
- 🔍 AI-based analysis of MAC data — MAC already pools intelligence from central and state agencies. Shah wants that pooled data run through AI tools to surface patterns faster.
- 📋 SOPs built from that analysis — the output of the AI review is meant to feed into new Standard Operating Procedures for the agencies, not the other way around.
- 🎯 A parallel push on PRAHAAR — Shah also emphasised training for MHA’s PRAHAAR scheme, along with faster extradition of fugitives and a more integrated approach to UAPA investigations.
- 🗂️ Broader conference agenda — the two days also covered narcotics control, illegal migrant deportation, cybercrime, proxy warfare, blue-economy security, and even a proposed bio-security framework.
Nothing here names a specific AI vendor, model, or platform. No budget figure was announced. No procurement timeline was given.
Why It’s Still a Big Deal
Even without the specifics, the direction of travel matters. MAC is the backbone that’s supposed to make India’s dozens of intelligence and law-enforcement agencies talk to each other — and this isn’t happening in isolation. State police forces have already started piloting their own AI investigation tools this year, like Gujarat Police’s NARIT system for NDPS cases, suggesting the shift toward AI-driven policing is underway at the state level too.
Telling it to run AI-based analysis is a signal that India’s security apparatus is moving from “collect and store” intelligence toward “collect, analyse, and act” — the same shift Western agencies like the NSA and MI5 have been making for years with predictive analytics and pattern-recognition tools.
For India’s AI industry, this is worth watching closely — but it’s important not to get ahead of the story. Nothing announced this week is a tender, a contract, or a budget line. What it does signal is a policy direction: if MAC is genuinely scaling up AI-based intelligence fusion, the demand for secure data-analysis platforms, natural language processing for regional-language intelligence, and defence-grade infrastructure will likely follow — eventually, through official channels, not overnight.
The India Angle That Matters
This is where it gets interesting for Indian developers and AI companies specifically. Most government AI conversations in India this year — the IndiaAI Mission, the GPU compute subsidy scheme, the India AI Impact Summit — have centred on economic growth, healthcare, and agriculture. Security and defence AI has stayed comparatively quiet in public discourse, even though agencies have clearly been experimenting with it since at least mid-2025.
Shah’s comments put a government voice, on record, behind the idea that India’s counter-terrorism apparatus should be AI-native. For a country trying to build a “frugal, sovereign, and scalable” AI ecosystem — Shah’s own ministry colleague Ashwini Vaishnaw used almost those exact words at the India AI Impact Summit in February — this is the first time that sovereignty language has been tied explicitly to national security use cases rather than just compute infrastructure.
What to Watch Next
- 🏛️ Whether MHA follows up with any formal procurement process or AI policy framework specific to MAC
- 🔗 Whether this connects to the IndiaAI Mission’s compute allocation, or stays a separate track entirely
- 📅 How Shah’s push at NSSC-2026 compares with what actually gets implemented by NSSC-2027 — past years’ directives on this exact topic haven’t always translated into visible on-ground change
Bottom Line
Read the coverage carefully and this is less a single dramatic AI announcement and more the latest checkpoint in a policy push that’s been building quietly since mid-2025. It’s real, it’s official, and it does matter for anyone tracking where India’s AI and defence-tech industries are headed — but it isn’t (yet) the tender-generating gold rush some of the framing suggests.
FAQs
What did Amit Shah actually direct security agencies to do?
He directed agencies to run AI-based analysis on intelligence available through the Multi-Agency Centre (MAC) platform, and to use that analysis to build new SOPs for national security challenges.
Does this mean India is launching a new defence-AI tender or budget?
No. No vendor, budget, or procurement timeline was announced. This was a policy directive at a conference, not a funding or tender announcement.
Is this a brand-new AI policy for India’s security agencies?
Not entirely — Shah gave a similar directive on using AI and Big Data at MAC back in July 2025. This week’s statement builds on that, rather than starting from scratch.