AI, M-governance and social media: governing the digital frontier

The same phone that receives a farmer’s subsidy through mobile authentication can now be used to fool voters with a deepfake video of a candidate who never said those words. That tension, between the state’s embrace of artificial intelligence for service delivery and its scramble to contain what AI lets citizens and criminals do, is the defining governance story of the decade. India has chosen a distinctive path: build the AI infrastructure first, keep regulation light, and let the IT Act, the DPDP Act and sectoral rules do the early running. This article maps that choice.

It is best read alongside the e-governance article (governance-06-e-governance), which covers the DPI stack, and the data-protection article (governance-07-data-protection), which covers the DPDP Act in depth. Here the focus is narrower: the IndiaAI Mission and its GPU strategy, AI inside government work like policing and grievance redressal, the deepfake and cyber-fraud emergency, M-governance as the citizen’s pocket government, and the unfinished argument over how to regulate social media, including whether minors should be kept off it.

The IndiaAI Mission: compute as sovereignty

The working definition first:

The IndiaAI Mission, approved in March 2024 with an outlay of over ₹10,300 crore, treats compute as a strategic asset. Its compute pillar has onboarded more than 38,000 GPUs offered at a subsidised ₹65 per hour, answering the chronic complaint that Indian startups and researchers depend on foreign cloud providers. Alongside sit an Application Development Initiative (30 applications approved by July 2025), AIKosh, the dataset platform with over 5,500 datasets and 251 AI models across 20 sectors, a Foundation Models programme backing 12 startups to build large multimodal models on Indian data and languages, FutureSkills for researchers, startup financing, and a Safe and Trusted AI track that has selected 13 projects.

The mission did not arrive from nowhere. NITI Aayog’s National Strategy for AI (2018) set the “AI for All” vision and identified five focus sectors: healthcare, agriculture, education, smart cities and smart mobility. The mission is the hardware-and-money answer to that strategy. The geopolitical context matters for examiners: the Global South argument is that whoever owns the compute and the datasets sets the defaults of AI governance, so India is trying to buy a seat at the table rather than rent one.

The national AI initiatives, each with its one-line definition:

AI inside the state: policing, courts, farms, grievances

AI’s governance footprint is already concrete. In policing, the Indian Cyber Crime Coordination Centre froze 24.67 lakh mule accounts using AI-based fraud detection in 2024, and predictive-policing tools map crime hotspots. In the judiciary, the SUVAS tool translates legal documents into 22 regional languages for 10 high courts. In agriculture, the Kisan e-Mitra chatbot works in 11 languages, and AI advisories cover weather, pests and crop choice. In grievance redressal, AI routes and analyses complaints to cut pendency, while the Bhashini platform brings government services into more than 20 Indian languages.

The official summary claims are large: targeted welfare delivery with fewer leakages, personalised learning, faster disaster forecasting, and AI-synced traffic signals in cities like Bengaluru. The UPSC-relevant caution, visible across the source material, is that each of these is a high-stakes allocation decision disguised as a technical one. When an algorithm decides which families validate for PM-Kisan, and 2.18 crore families get stranded by validation errors, the governance question is not whether AI works but who audits it when it fails.

The structural challenges, in the standard framing:

The other face: deepfakes, fraud and the black box

Generative AI’s risks have arrived faster than its safeguards. Hyper-realistic deepfakes threaten elections, reputations and communal calm: the Rashmika Mandanna deepfake of November 2023 was the moment the issue went mainstream in India, and “digital arrest” scams now use cloned voices to frighten victims into transferring money. Indians lost ₹22,845 crore to cyber fraud in 2024, and CERT-In (Indian Computer Emergency Response Team) reported 2.94 million cyber incidents in 2025.

The structural problems go deeper. Algorithmic bias means models trained on skewed data discriminate against marginalised groups, denying them fair access to benefits. The black-box problem, the lack of explainability in deep learning, makes AI decisions opaque and accountability hard to assign. Add job displacement without reskilling, dependence on foreign GPUs, and the constitutional unease around facial-recognition surveillance, and the risk ledger is as full as the opportunity one.

The definition to memorise:

The statutory provisions invoked against deepfakes:

Beyond AI: blockchain pilots, drones and critical infrastructure

A blockchain is a shared, tamper-evident ledger: records are grouped into blocks, cryptographically chained, and replicated across nodes so no single official can alter them quietly. India has run it as a governance instrument rather than a currency play. Pilots include land-registration records in Telangana and Maharashtra, caste and income certificates in Karnataka, the Tamil Nadu Blockchain Backbone as a state-level shared ledger, an NFT-based land-mutation pilot in West Bengal, a grievance-redressal pilot in Firozabad, Uttar Pradesh, TRAI's spam-control framework and SEBI's KYC records. The governance logic is auditability: certificates and land records stop being documents an official can rewrite.

The Drone Rules, 2021 liberalised unmanned aircraft as a governance instrument: no pilot licence for micro drones, sharply reduced fees, and clearances through the Digital Sky platform. The governance use case is survey and mapping: the SVAMITVA scheme's drone mapping of villages for rural property cards is the flagship, with GIS layers feeding PM Gati Shakti planning and disaster mapping alongside it.

The cyber-institution stack guards all of this. CERT-In handles incident response; the National Critical Information Infrastructure Protection Centre, NCIIPC set up in January 2014 under the National Technical Research Organisation (NTRO) under Section 70A of the IT Act, 2000, protects critical information infrastructure in sectors like power, banking, telecom, transport and strategic government enterprises; the Indian Cyber Crime Coordination Centre, I4C (under the Home Ministry) coordinates cybercrime enforcement. The standing weakness: the National Cyber Security Policy has not been updated since 2013.

India’s “lightweight” AI regulation

India has deliberately chosen not to enact a dedicated AI law. Instead it runs a lightweight, principle-based approach: the IT Act of 2000, the DPDP Act of 2023, the Bharatiya Nyaya Sanhita of 2023 and the Consumer Protection Act of 2019 are stretched to cover AI, with sectoral regulators filling the gaps (RBI for fintech AI, SEBI for trading algorithms, IRDAI for insurance underwriting, ICMR’s ethical guidelines for biomedical AI). Contrast this with the EU AI Act (2024), which sorts AI into four risk tiers from unacceptable to minimal with penalties up to €35 million or 7% of global turnover, and China’s 2023 generative-AI rules with mandatory security assessments and algorithm registration. India has also led multilaterally: it chaired the Global Partnership on AI (GPAI) and hosted its summit in New Delhi.

The deepfake-specific machinery is the IT Amendment Rules of 2026, which mandate prominent labelling of Synthetically Generated Information (SGI) and require harmful deepfakes to be taken down within 3 hours. A Prof. Balaraman Ravindran committee drafted the India AI Governance Guidelines on Seven Sutras (trust as foundation, people first, innovation over restraint, fairness and equity, accountability, understandable by design, and safety, resilience and sustainability) and six pillars, proposing three new institutions: an AI Governance Group for coordination, an AI Safety Institute for testing, and a Technology and Policy Expert Committee for advice.

One proposal keeps recurring in the literature but has no explanatory home, so here it is, marked proposed, not enacted: the Digital India Act, which would replace the 25-year-old Information Technology Act, 2000. Its draft framework, which drew about 23,000 public suggestions, has three headline elements: redefine intermediary so it covers generative-AI systems, clarify the Section 79 safe harbour, and reframe platform duties for the AI era. Until Parliament passes it, the IT Act, 2000 remains the law.

M-governance: the pocket government

M-governance, a sub-domain of e-governance, uses mobile technologies (SMS, apps, USSD, IVRS) to deliver services anytime, anywhere, and it is where most citizens actually meet the digital state: about 96.6% of Indian internet users access the web through mobile devices. The UMANG app alone hosts over 1,745 services for 5.58 crore citizens. The highlights of the mobile state include the SACHET app, which sent over 100 million geo-fenced alerts during the 2025 monsoon, DBT through mobile authentication, the Kisan Suvidha app for market prices, and Udyam registration for MSMEs in under 3 minutes on a phone.

The problems are those of any pocket government built for the median smartphone owner. A 2025 NSO report found only 35% of rural adults can use a mobile app independently; advanced services are often English-centric; low-end phones lack storage for ten different government apps; rural networks remain unstable; and mobile apps are honeypots for cyber criminals. The reform menu in the sources is instructive: super-apps to end fragmentation, voice-first governance through Bhashini, offline modes that sync when the network returns, zero-rated data for essential portals, and Estonia’s once-only principle, where the citizen gives her data once and departments move it between themselves.

Regulating social media: speech, traceability and minors

The working definition:

Social media is both a governance instrument and a governance problem. MyGov, launched in 2014 with over 3.5 crore registered users, is among the world’s largest citizen-engagement platforms, and the Pre-Legislative Consultation Policy of 2014 pushes draft bills through digital channels. But viral misinformation during elections and communal incidents has had real-world consequences, including lynchings, while WhatsApp mobilisation in regional languages stays hard to monitor.

The positive ledger examiners expect:

The regulatory case:

The regulatory spine is the IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021: intermediaries must publish content-moderation policies, appoint a Chief Compliance Officer, a Nodal Contact Person and a Resident Grievance Officer, and respond to takedown orders within fixed timelines. Significant Social Media Intermediaries with over 50 lakh users face extra duties, including first-originator traceability for serious offences. The 2023 amendment’s Fact Check Unit was struck down by the Bombay High Court in September 2024 as unconstitutional government overreach over free speech, a reminder that Shreya Singhal v. Union of India (2015) remains the constitutional benchmark. The October 2025 and February 2026 amendments layered the synthetic-information regime described above.

The newest front is the minors debate. Karnataka announced plans to ban social media for under-16s and Andhra Pradesh announced restrictions for under-13s; Australia’s Online Safety Amendment Act of 2024 was the world’s first under-16 ban, with penalties up to AUD 49.5 million, and Indonesia became the first Southeast Asian nation to enforce one in March 2026. The case for bans cites the Economic Survey 2025-26’s point that adolescents lack the “neurological brakes” for infinite scroll, plus mental-health harm, cyberbullying and data-harvesting without informed consent. The case against cites fake-ID circumvention, privacy risks of age verification, free-speech concerns, and the risk of pushing teens underground. The emerging consensus points to a tiered model: strict restrictions under 13, supervised access for 13 to 15, full access at 16, with safe-by-design defaults and algorithmic transparency.

OTT platforms and the Digital Media Ethics Code

Over-the-Top (

The case for regulation has four planks: the

The government's answer is the

What examiners keep asking

• The inadequacies that hamper e-governance’s effectiveness, transparency and accountability (2023).
• How digital illiteracy and poor ICT accessibility hinder socio-economic development (2021).
• The vital factors behind ICT project implementation failures (2019).
• The Fourth Industrial Revolution’s impact on e-governance (2020).

Maps from space and money on a ledger: Bhuvan and the Digital Rupee

Bhuvan is ISRO’s geo-platform, a portal that gives government departments satellite imagery, maps and geospatial services for planning. It underpins the layered GIS maps of PM Gati Shakti and the drone-based village surveys of SVAMITVA, which has mapped more than 1.7 lakh villages for property cards. A geo-platform matters because planning moves from outdated paper maps to current, layered spatial data: the same road project can be checked against forest cover, land records and flood zones on one screen.

Money is being tested on digital rails too. The Digital Rupee, the RBI’s central bank digital currency (CBDC) pilot, tests a sovereign digital alternative to cash and private crypto. Unlike UPI, which is a payment rail that moves bank money, a CBDC is actual central-bank money in digital form. Its governance significance is control: the state gets a digital payments instrument it issues, monitors and settles itself, without depending on private networks.

The same data-first instinct is spreading inside departments. Indian Railways uses the Ideal Train Profile, an AI tool, to plan train composition and utilisation, and NITI Aayog has built cancer image databases to train diagnostic AI. Each is a small version of the larger argument of this article: the state is learning to govern with data first and paperwork second.

Key Terms

• Indian Cyber Crime Coordination Centre, I4C (under the Home Ministry: The Indian Cyber Crime Coordination Centre (I4C) is an attached office of the Union Home Ministry that coordinates the national response to cybercrime. Its scheme was approved in 2018 and it was inaugurated in 2020. It runs the National Cyber Crime Reporting Portal (cybercrime.gov.in), the 1930 helpline, and the Citizen Financial Cyber Fraud Reporting and Management System for freezing defrauded funds. For UPSC, I4C is the key institution for GS-III questions on cyber security and digital fraud. The 1930 toll-free helpline, whose linked fraud-reporting system has saved thousands of crores in defrauded funds.
• SVAMITVA scheme's drone mapping of villages for rural property cards: SVAMITVA (Survey of Villages and Mapping with Improvised Technology in Village Areas) is a Ministry of Panchayati Raj scheme launched on 24 April 2020 that uses drones to survey inhabited rural areas and issue property cards to household owners. With the Survey of India as technology partner, it creates accurate land records, enables property monetization and reduces disputes. It matters for UPSC because drone-based cadastral mapping links land reforms, digital governance and rural empowerment. National Panchayati Raj Day, 24 April 2020
• National Critical Information Infrastructure Protection Centre, NCIIPC: The National Critical Information Infrastructure Protection Centre (NCIIPC) is India's designated agency for protecting critical information infrastructure, created under Section 70A of the Information Technology Act, 2000 (inserted by the 2008 amendment). It identifies critical sectors, issues threat advisories, and coordinates cyber-incident response for systems whose disruption would cripple national security or the economy. It matters for UPSC because cybersecurity, critical infrastructure, and digital governance are key GS-3 internal-security topics. Section 70A of the Information Technology Act, 2000
• National Pest Surveillance System: The National Pest Surveillance System uses AI-driven predictive insights for pest control and weather in agriculture, a standard example of AI inside the state.
• Tamil Nadu Blockchain Backbone: Tamil Nadu Blockchain Backbone is the state's blockchain-as-a-service infrastructure, branded Nambikkai Inaiyam (meaning trusted link), launched by TNeGA in June 2023 with the e-Pettagam citizen wallet app. It secures e-Sevai certificates, academic records, and land transaction data against tampering through blockchain verification. For UPSC, it is a GS-2/GS-3 e-governance case study on using emerging technology for trustworthy digital public infrastructure. Nambikkai Inaiyam, launched June 2023
• Consumer Protection Act 2019: The Consumer Protection Act, 2019 replaced the 1986 Act to modernise consumer rights for the digital age. It established the Central Consumer Protection Authority with powers against unfair trade practices, covered e-commerce transactions and endorsements, introduced product liability claims, and created District, State, and National Commissions with higher pecuniary limits. It is central to questions on consumer rights and market regulation. A buyer of a defective phone can now claim product liability directly from the manufacturer.
• Online Safety Amendment Act: The Online Safety Amendment Act is Australia's Online Safety Amendment (Social Media Minimum Age) Act 2024, which bars children under 16 from holding accounts on major social media platforms. In force from December 2025, it requires platforms to take reasonable steps to prevent underage accounts, with heavy penalties for failure, and it is being challenged in Australia's High Court. For UPSC it is the world's most cited experiment in age-gating social media and a reference point for comparative digital regulation. Platforms such as Instagram and TikTok had to introduce age-assurance checks for Australian users after the Act took effect.
• Digital Media Ethics Code: The Digital Media Ethics Code requires OTT classification into U, U/A 7+, U/A 13+, U/A 16+ and A, with parental locks and age verification, Press Council and Cable TV norms for digital news, and a three-level grievance redressal ending in an Inter-Departmental Committee.
• National Strategy for AI: The National Strategy for AI is NITI Aayog's 2018 framework, formulated to leverage AI for socio-economic growth while addressing challenges and maximizing opportunities. It set the "AI for All" vision across healthcare, agriculture, education, smart cities and smart mobility.
• Section 79 safe harbour: Section 79 of the Information Technology Act, 2000 grants intermediaries such as social media platforms and ISPs 'safe harbour', shielding them from liability for third-party content if they act as neutral conduits, observe due diligence and take down unlawful content on government or court notice. Upheld with a reading down in Shreya Singhal (2015), it underpins the 2021 IT Rules. For UPSC, it is core to GS-2/GS-3 debates on free speech, platform accountability and digital regulation. Shreya Singhal v. Union of India (2015)
• TRAI's spam-control framework: TRAI's spam-control framework is the Telecom Regulatory Authority of India's system for curbing unsolicited commercial communication (UCC). It uses blockchain-based distributed ledger technology (DLT) to register telemarketers and verify message headers and templates, with 2024 amendments requiring whitelisting of URLs and links in commercial SMS to fight fraud. For UPSC, it is a GS-2/GS-3 current-affairs case of technology-based regulation and consumer protection in digital governance. 2024 UCC amendments (DLT whitelisting)
• Telangana and Maharashtra: The pairing points to the Godavari water-sharing relationship between Telangana and Maharashtra, centered on barrages over the Godavari and its tributaries the Pranahita and Penganga. The two states signed a pact in 2016 for three barrages, though the Tummidihatti barrage height remains contested. For UPSC, it illustrates cooperative federalism and inter-state river dispute resolution. Tummidihatti barrage (disputed project on the Pranahita)

Practice questions