MahaAI: Building Safe, Secure & Smart Governance

Contents

Executive Summary

India is positioning itself as a leader in responsible AI governance through the inaugural AI Impact Summit 2026, hosted in Maharashtra. The summit emphasizes that governance must actively shape AI rather than be shaped by it—requiring intelligent, human-centered policies that balance innovation with safety, establish global cooperation frameworks, and address emerging threats like deepfakes, quantum computing vulnerabilities, and algorithmic bias.

Key Takeaways

AI governance success depends on policy wisdom, not algorithmic sophistication. History will judge governments by how well they governed intelligence, not by technological sophistication. India has an opportunity to model responsible AI governance in the Global South.
Data sovereignty and local economic benefit must be protected proactively. Without active governance, India's abundant datasets (health, demographic, transactional) will be monetized by foreign companies without reciprocal benefit. State data authorities and careful licensing frameworks are essential.
Deepfakes and quantum computing threats require immediate, coordinated response. Detection of synthetic media is an active research challenge; quantum computing will break current encryption within years. India must invest in quantum-safe cryptography and hybrid verification ecosystems now.
Infrastructure (broadband, electricity, education) is the real bottleneck for scaling AI benefits. Tier 2/3 cities cannot benefit from AI governance innovations without foundational improvements to internet speed, power reliability, and workforce literacy—not technology alone.
Ethics, bias, and harm prevention must be embedded before deployment, not reviewed after. Smart glasses, facial recognition, and other surveillance-enabling technologies need pre-market governance review by independent institutes (e.g., India Safety Institute) to prevent harm to vulnerable populations.

Key Topics Covered

Governance Philosophy & Policy Framework: The governance paradox (regulate too slowly vs. too heavily); principles of human-centered design, transparency, accountability, risk-based regulation, and adaptive policies
Maharashtra's MahaAI Initiative: State-level AI governance infrastructure including crime prevention (Mahak Crime OS), intelligent government services, and digital sovereignty
Cybersecurity & Law Enforcement: AI applications in detecting cyber crimes, preventing fraud, and preparing for emerging threats (quantum computing, cyber warfare)
Deepfakes & Synthetic Media Detection: Research challenges in detecting AI-generated audio, video, and images; federated learning and encrypted detection approaches
Digital Public Infrastructure (DPI) & AI Integration: Population-scale digital rails (Aadhaar, UPI, DigiLocker) as foundations for AI-enabled precision governance and dynamic eligibility assessment
Enterprise & Government Collaboration: Large technology companies (TCS) supporting state governance through common databases, platform intelligence, and integrated service delivery
Ethical AI & Gender Considerations: Bias, privacy concerns, device safety (e.g., smart glasses), and the need for guardrails before deploying new technologies
Scaling to Tier 2/3 Cities: Education, internet infrastructure, cost, and addressing dual challenges of technology adoption and preventing societal "dumbing down"
Economic & Employment Impact: Job displacement in traditional sectors vs. emerging opportunities; need for workforce reskilling at scale

Key Points & Insights

AI Governance is Strategic Infrastructure: Multiple panelists framed public data, cloud compute, and AI governance frameworks as essential infrastructure—comparable to energy, transport, and telecom—requiring coordinated state and national investment.
Data as Sovereign Asset: Maharashtra's State Data Authority is working to prevent monetization of India's data by foreign companies. Example given: pharmaceutical companies exploiting India's large population health datasets without reciprocal benefit. Governance must ensure data is used for national economic and social benefit, not extracted for free.
Deepfakes Require Multimodal Defense: Detecting AI-generated audio/video is computationally difficult in real-world conditions (noise, multiple languages, limited training data). Research is moving toward federated learning with differential privacy to enable collaborative detection without exposing proprietary models or sensitive data.
DPI as Enabler of Precision Governance: Verifiable digital credentials (moving from static identity to dynamic eligibility) allow AI to predict who needs subsidies or benefits before crisis occurs. This requires guardrails: explainability, auditability, and human redressal pathways to prevent algorithmic harm.
Quantum Computing Is an Immediate Governance Threat: Current RSA encryption, blockchain, and banking systems are vulnerable. China has invested $15–20B; India has invested $1B. Urgent need to transition to quantum-resistant cryptography and prepare infrastructure before quantum computers mature.
Ethics Cannot Be Bolted On: Devices like smart glasses pose unaddressed privacy and consent risks that regulatory frameworks have not resolved. Technologies should be contextualized to local values before deployment; India should not blindly adopt solutions designed for other governance contexts.
Skill & Infrastructure Deficits Limit Scale: Only 20% of Maharashtra's 9-crore workforce is at skill levels 3–4; 80% are at levels 1–2. Internet penetration and quality are inadequate even in tier-1 cities (Mumbai: 58 Mbps average). Tier 2/3 cities face acute infrastructure gaps. Education reform must precede technology scaling.
AI-Powered Government Services Show Immediate ROI: Maharashtra Police's cyber helpline (1930) froze ₹1,000+ crore in scam proceeds and prevented ~70 suicide attempts from cyber bullying in less than 6 months—demonstrating tangible public safety impact.
Human-Centered Design is Non-Negotiable: Panelists consistently stressed that AI should "scale empathy through insight" (make governance faster, more responsive, more inclusive) rather than distance citizens from the state. Explainability and human oversight must remain central.
Global Coordination on AI Standards is Essential: Since AI "does not recognize borders," interoperable frameworks, shared safety standards, and cooperative oversight mechanisms are necessary—requiring alignment beyond national policies.

Notable Quotes or Statements

"The question before us is not whether AI will shape governance. The question is whether governance is going to shape artificial intelligence." — Opening speaker

"History will not judge us by our sophisticated algorithms. It will judge us by the wisdom of our governance." — Opening speaker

"The industrial revolution reshaped economies. The digital revolution reshaped communication. The AI revolution will reshape decision making itself. With that power comes great responsibility." — Opening speaker

"Use AI not to distance the state from the citizens but to make governance more human, faster, more responsive and more inclusive. In other words, scale empathy through insight." — Shri Ashish Shelar, Minister of IT & Cultural Affairs, Government of Maharashtra

"In less than six months more than 1,000 crores of rupees which would have gone into the hand of the scamsters have been frozen and are being ultimately returned to the victims... and more than 70 young girls who were being subjected to intense cyber bullying have been saved." — Shri Yashi Yadav, ADG Cyber Police, Maharashtra

"AI is not a technical tool which is fundamentally going to make the governance more efficient. It is fundamentally meant for how to bring the benefit welfare and happiness to the community at large." — Major Ranjit Gowami, Head Corporate Affairs, Tata Consultancy Services

"What we sometimes seem to miss is the wood for the trees... when you're building out devices that endanger half the population, that's where it breaks down." — Bina Sarkcar, on ethical AI and gender considerations

"AI is also going to be the biggest dumbing down element for the society... How do we set our education system right?" — Dr. Amit Kapoor, Chair, Institute for Competitiveness

"Let us govern intelligence with wisdom." — Opening speaker (closing sentiment)

Speakers & Organizations Mentioned

Government Officials

Shri Ashish Shelar – Minister of IT & Cultural Affairs, Government of Maharashtra
Shri Pravin Pardeshi – CEO, MITRA (Maharashtra IT & Tech Services); Chief Economic Adviser to the Chief Minister
Shri Yashi Yadav – Additional Director General of Police, Maharashtra Cyber Department
Chief Minister Devendra Fadnavis – Government of Maharashtra (referenced throughout)

Academic & Research Institutions

Dr. Anupam Chhatarpadhya – Associate Professor, Nanyang Technological University, Singapore (deepfakes, quantum AI)
Dr. Danesh – IIT Mumbai (working on Maha GPT, government order analysis)
Dr. Amit Kapoor – Chair, Institute for Competitiveness (on tier 2/3 scaling challenges)
Dr. Ganesh Ramakrishnan – (mentioned, affiliation not fully specified)

Private Sector & Industry

Major Ranjit Gowami – Head Corporate Affairs, Tata Consultancy Services (TCS)
Suresh Sati – Managing Director & CEO, Protein Ego of Technologies (DPI/AI integration)
Bina Sarkcar – Customer Success Executive, ServiceNow; Volunteer, Women for Ethical AI South Asia (UNESCO-powered)
Microsoft – Satya Nadella referenced for Mahak Crime OS collaboration

Government Agencies & Initiatives

MITRA – Maharashtra IT & Tech Services (State AI governance)
State Data Authority – Maharashtra (data sovereignty)
Cyber Helpline 1930 – Maharashtra Police
India Safety Institute – 2025 establishment (device safety governance)
NITI Aayog – Referenced for employment analysis
Mahak Crime OS – AI-powered crime prevention system (Maharashtra Police)
Maha GPT – Government order analysis system (in development)

Technical Concepts & Resources

AI & Governance Systems

Decision Intelligence: AI applied to government decision-making at scale
Mahak Crime OS: Microsoft-backed AI system for crime prevention, detection, and investigation
Maha GPT: Small language model (not large LLM) for querying government orders and regulations at citizen and officer levels
Intelligent Government Infrastructure: Cloud-native, modular, API-driven backbone integrating AI into public services

Data & Infrastructure

DPI (Digital Public Infrastructure): Population-scale systems including:
- Aadhaar – Identity layer
- UPI – Payment/transactional layer
- DigiLocker – Document storage with millions of authenticated documents
Verifiable Digital Credentials: Machine-readable, cryptographically verifiable attributes enabling dynamic eligibility
Blue Dot Concept: Machine-readable attributes of individuals enabling AI-driven precision governance

Deepfakes & Synthetic Media Detection

Federated Learning: Multiple distributed models merged without exposing training data or model weights
Differential Privacy: Technique adding "wrapper" privacy to model merging, preventing data inference
Fully Homomorphic Encryption: Enabling AI operations on encrypted data (noted as slow/computationally intensive)
Synthetic Data Generation: Creating realistic training datasets when labeled real-world data is scarce
Truth Checker Mechanism: Cross-referencing detected deepfakes against news/media coverage to ground reality assessment
Noise Robustness: Training models on progressively noisier samples to improve real-world detection accuracy

Cybersecurity & Threat Intelligence

Dark Web Monitoring: Surveillance of underground internet for threat actors, stolen data
Threat Intelligence Tools: Luminar, Cognite, Pathfinder (big-data analytics for cyber threats)
Quantum Computing Threats:
- Can solve complex problems in <6 seconds (supercomputers: 50+ years)
- RSA encryption vulnerability: Quantum computers can break current banking/blockchain encryption
- Nation-state investment: China $15–20B; India $1B
- Risk to: Bitcoin, credit card systems, blockchain, banking infrastructure

AI Governance Frameworks

Five Pillars of Smart Governance (Maharashtra):
1. Compute and cloud at scale
2. High-quality public datasets
3. State AI governance
4. Interoperability and standards
5. Capacity building
Internet Health as Policy: Treating digital health equivalent to physical health; addressing disinformation, deepfakes, cyber attacks
Hybrid Verification Ecosystems: Combined robust cyber security + digital literacy + critical thinking

Employment & Skills

Skill Levels 1–4 Distribution: Maharashtra workforce ~80% at levels 1–2 (low skill); ~20% at levels 3–4 (high skill)
Post-2020 Job Trends: Per NITI Aayog analysis—post-graduate engineers showing 95%+ employment (pre-2020), but declining to lower rates post-2020; blue-collar workers (masons, caregivers) showing 65% rising employability trend
Workforce Underemployment: Maharashtra ~50% underemployed despite skill development initiatives

Maharashtra GDP Contribution: 17–18% of India's total; termed "engine of growth"
IT Workforce Concentration: ~16% of India's tech workforce concentrated in Pune (single city)
Malnutrition in Maharashtra: ~50% of population malnourished despite state development

Additional Context & Significance

AI Impact Summit 2026: First global AI summit hosted in the Global South; ~20 heads of state, 60 ministers, hundreds of AI leaders attending
MahaAI Initiative: Positioned as a "living laboratory" for AI in governance, showcasing practical deployments in crime prevention, public services, disaster response, and welfare delivery
Global South Perspective: India explicitly framing itself as writing the "operating system of the AI age" for developing nations, modeling responsible AI governance distinct from Global North approaches
Quantum Computing Timeline: Urgent threat horizon—encryption vulnerable "within years," not decades

This summary reflects the transcript as provided. Specific technical claims (e.g., quantum computing timelines, employment statistics) are sourced directly from panelist statements and should be verified against peer-reviewed research for policy application.