Intelligent Telecom Networks in the age of AI and Data Sovereignty
Detailed Summary
- Purpose: Set the stage for the convergence of AI, telecom, and data sovereignty within India’s DPI agenda.
- Key Points
- GSMA unites the mobile economy to unlock connectivity that fuels industry and society.
- India is at a pivotal juncture: long‑standing DPI (identity, payments, commerce) now being reshaped by AI, real‑time data, and autonomous systems.
- Networks are transitioning from passive pipes to intelligent, programmable layers that influence AI model performance, edge optimisation, fraud prevention, and digital‑identity security.
- Digital sovereignty now means control over infrastructure, standards, and the intelligence that powers national systems, not merely data residency.
- The session’s goal: move from high‑level themes to tangible actions, collaborations, and learning from India’s experience.
2. Panelist 1 – Rahul Vatts (Bharti Airtel)
2.1 The Network as Trusted Infrastructure
- Scale of Airtel’s backbone: > 1 million BTS sites, > 5 million km of fiber, > 1,000 edge/hyperscale data centres, each Mobile Switching Centre serving 30‑50 million users.
- Economic impact: January 2024 alone, ₹28 lakh crore processed via UPI across a billion‑plus users – all underpinned by the telecom network.
2.2 Trust‑by‑Design: OTP, SMS, and Aadhaar‑Enabled Payments
- OTP/SMS messages act as real‑time trust signals (≤ 2 ms latency).
- Aadhaar‑linked payments: > ₹500 million in a single transaction cohort, facilitated by ultra‑low‑latency connectivity.
2.3 AI‑Driven Financial Services (Digital Lending & Fraud Scoring)
- Airtel provides telco‑derived risk indicators to banks, enabling instant, micro‑loan decisions (loans ≤ ₹2 lakh).
- Fraud‑prevention products:
– Spam‑call warning,
– Blocking malicious links (partnered with Google),
– Real‑time “friction” prompts during OTP‑related calls.
2.4 Exporting Indian DPI Blueprints
- Airtel’s DPI Inbox solution is being adapted for African markets, offering hardware, software, and capped cloud to replicate India’s stack abroad.
Key Insight: The telco layer is not a passive conduit; it creates trust, enriches financial inclusion, and serves as a growth engine for public‑sector services.
3. Panelist 2 – Martin (interpreted as Ambika Khurana, Vodafone Idea)
3.1 Avoiding Parallel Digital Infrastructures
- Concern: Duplication of capabilities when Mobile Network Operators (MNOs) build “operator‑led” APIs that overlap with government‑run DPI services.
3.2 Four Pillars for DPI Evolution
- Enrichment – Adding contextual data to raw telecom signals.
- Serviceability – Ensuring DPI reaches the last‑mile civilian and supports financial inclusion.
- Purpose‑Driven Architecture – DPI must serve both social (inclusion) and economic (payments) goals.
3.3 Converged Platform Experience
- Described the Digital Intelligence Platform (DIP) that aggregates TSP (Telecom Service Provider) data, exposing enriched context (e.g., simultaneous location vs. call activity) to banks and authentication entities.
3.4 Open APIs & Collaboration
- Cited collaboration with GSMA OpenGate APIs and the role of DOT (Department of Telecommunications) in standardising interfaces.
Key Insight: A converged, open‑API ecosystem prevents siloed solutions, enabling richer, real‑time data for fraud detection and credit decisions while preserving DPI’s social mission.
4. Panelist 3 – Deepak Maheshwari (CSEP)
4.1 Re‑defining Data Sovereignty in an AI‑Driven DPI Era
| Aspect | Traditional View | Maheshwari’s Expanded View |
|---|---|---|
| Physical localisation | Data stored within national borders. | Still essential for defence/finance, but insufficient alone. |
| Contextual localisation | Language‑specific interfaces. | Need local relevance (e.g., weather forecasts specific to locale). |
| Agency of citizens | Users may wish to share data internationally (e.g., visa applications). | Recognise user‑controlled data mobility while protecting national interests. |
| Economic participation | India as a global data‑processing hub. | Balance two‑way data flows – avoid “walls” that block inbound data needed for AI. |
4.2 Sovereignty as Participation, Not Control
- Emphasises contributing to global standards (GSMA, ISO, ITU, IEEE) rather than attempting unilateral control.
- Calls for institutional mechanisms that provide India a greater “share of the pie” when standards are set.
4.3 Historical Perspective
- India’s telecom heritage dates back to 1854 telegraph act and 1858 submarine cable – the “binary world” of the past.
4.4 The 3C Framework (Carriage, Content, Conduct)
- Carriage – Network infrastructure (must exist).
- Content – Data flowing through (requires governance).
- Conduct – Policies, regulations, and ethical guidelines (critical for AI).
Key Insight: Data sovereignty in the AI age is multidimensional: physical control, contextual relevance, citizen agency, and active participation in standard‑setting.
5. Panelist 4 – Mansi Kedia (World Bank)
5.1 Risks of Siloed Public & Private Digital Capabilities
- Trust erosion: Multiple independent systems increase vulnerability points.
- Efficiency loss: Fragmentation hampers economic gains from data sharing.
- Innovation slowdown: Closed ecosystems limit novel AI‑driven services (e.g., mobile‑data‑based credit scoring).
5.2 Standards vs. Blueprints
- Standards – Prescriptive, enable interoperable network operation.
- Blueprints – Flexible, adaptable guidelines that incorporate best practices across jurisdictions.
5.3 World Bank’s Role
- Developed a DPI & Development Report outlining DPI principles, objectives, and what is / is not DPI.
- Promotes AI Commons and global blueprints to help emerging economies adopt DPI without reinventing the wheel.
Key Insight: Global standards create trust and interoperability, while blueprints provide the context‑aware flexibility needed for diverse economies.
6. Panelist 5 – Rahul Vatts (Operator Perspective on Data Sovereignty)
6.1 Misconceptions about “Sovereign” Cloud
- Physical residency ≠ sovereignty – control planes often remain abroad, exposing data to foreign jurisdiction (e.g., US CLOUD ACT).
6.2 Four “Slices” of True Sovereignty
| Slice | Description |
|---|---|
| Data Residency | Data stored within national borders (basic requirement). |
| Control‑Plane Sovereignty | Ability to manage and orchestrate data/compute locally. |
| Operational Sovereignty | Software patches, network upgrades performed by domestic teams. |
| Jurisdictional Sovereignty | Immunity from extraterritorial legal demands (e.g., US CLOUD ACT). |
6.3 Airtel’s Sovereign Cloud Initiative
- Launched Airtel Sovereign Cloud (≈ 140 crore TPS capacity).
- Selective data strategy: Critical KYC, health, defence data stay in‑country; other workloads may leverage hyperscalers for efficiency.
6.4 Regulatory & Policy Needs
- Need dynamic regulations that recognise AI’s reliance on data control.
- Encourage industry‑wide sovereign‑cloud offerings while maintaining interoperability with global services.
Key Insight: True data sovereignty requires control over the full stack—storage, orchestration, operation, and legal jurisdiction—not just “data‑at‑rest” within borders.
7. Panelist 6 – Martin / Ambika (Regulatory Lens)
7.1 Policy Frictions as Networks Turn AI‑Driven
- Reasonable AI (as per PM) – demands explainability, accountability, and fairness.
- Existing Unified License (UL) framework covers telecom services but does not yet address AI‑specific obligations (e.g., explainability of automated fraud blocks).
7.2 Emerging Regulatory Gaps
| Gap | Example |
|---|---|
| AI Explainability | Blocked scam calls must justify the decision to avoid “arms race” with scammers. |
| Digital Intermediary Rules | When telcos monetize secondary data insights, do they become “digital intermediaries” under new privacy law? |
| Spam & Scam Protection | Need balanced standards that protect users without over‑restricting legitimate communications. |
7.3 Path Forward
- Develop playbooks/blueprints that translate global standards into nationally adaptable guidelines.
- Regulators should stay innovation‑friendly, allowing pilots and flexible interpretations while safeguarding privacy and security.
Key Insight: Regulatory frameworks must evolve in tandem with AI‑enabled network services, providing clear, adaptable standards without stifling innovation.
8. Global Perspective – India’s DPI Model for the Global South (Deepak Maheshwari)
8.1 Open, Non‑Monetised Architecture
- India’s DPI stack is open‑protocol, royalty‑free, allowing other nations to adapt without licensing fees.
8.2 Diplomatic Enablement
- Think‑tank support (Ministry of External Affairs, Indian Council of World Affairs) provides soft‑power outreach, capacity‑building, and policy‑exchange.
8.3 Standards & Interoperability
- Reference to World Development Report on Standards – showcases how simple, long‑standing standards (e.g., traffic‑light colours) become global norms.
8.4 Sustainability Lens (Equity, Ethics, Ecology)
- Emphasis on resource‑efficient AI (low carbon footprint, minimal material usage) and “lifestyle for environment” philosophy.
Key Insight: India can export a scalable, adaptable, and sustainable DPI blueprint that enables Global South nations to achieve digital sovereignty without technological lock‑in.
9. World Bank Outlook – Scaling India’s DPI (Mansi Kedia)
- Evidence‑based guidance: India’s massive, heterogeneous deployment offers concrete lessons on what works and what does not.
- Collaboration with international bodies (BIS, UN, GSMA) to extend fast‑payment systems (e.g., UPI) and Finternet concepts.
- Mobile‑data‑driven planning: DPI’s open data model is already being leveraged for urban mobility, credit scoring, and public‑service planning.
Key Insight: India’s DPI experience is shaping global development strategies, especially where mobile penetration is high and public‑service digitisation is a priority.
10. Audience Q & A (Highlights)
| Questioner | Theme | Summary of Response |
|---|---|---|
| Vijay Agarwal (jewellery manufacturer) | Physical‑hardware data vault (ring) | Suggests a personal data‑embassy ring storing encrypted KYC/medical data, accessible only with user‑consent and blockchain verification. Panel acknowledges feasibility but notes existing Aadhaar security and regulatory data‑share practices need tightening. |
| Unidentified participant | Data embassies (multilateral data‑sovereignty hubs) | Consensus that reciprocal data‑embassy arrangements could work if built on mutual trust and governed by multilateral agreements. |
| Various audience members | Government data‑request practices | Critique of mandatory collection of non‑essential personal data (e.g., marital status for IRCTC). Call for data‑minimalism in public‑sector onboarding. |
| Moderator (Devashish) | Closing remarks | Thanked all panelists and audience; reiterated need for collaborative standards, sovereign‑cloud offerings, and continued dialogue. |
Key Takeaways
- Intelligent networks are now the backbone of DPI – they embed AI, enable real‑time fraud detection, and provide the trust layer for massive financial transactions (₹28 lakh crore in Jan 2024).
- Scale matters: Airtel’s infrastructure (1 M+ BTS, 5 M km fiber, 1 000+ edge data centres) supports hundreds of millions of concurrent users and underpins critical services like OTP, Aadhaar payments, and micro‑lending.
- Data sovereignty ≠ mere localisation – it encompasses control‑plane ownership, operational autonomy, jurisdictional immunity, and participation in global standards‑making.
- Open APIs and converged platforms (e.g., GSMA OpenGate, Digital Intelligence Platform) prevent fragmented “parallel” DPI layers and unlock richer, context‑aware services for banks and authentication entities.
- Standards vs. blueprints: Standards ensure interoperability; blueprints provide adaptable, context‑sensitive guidance for emerging economies.
- Regulatory evolution is urgent: Existing telecom licences (Unified License) don’t fully cover AI explainability, digital‑intermediary responsibilities, or spam‑scam mitigation; regulators must issue flexible playbooks.
- India’s DPI model is export‑ready: Open, royalty‑free protocols and diplomatic capacity‑building enable other countries, especially in the Global South, to adopt the stack without vendor lock‑in.
- Sovereign cloud is a reality, not a myth: Airtel’s own sovereign‑cloud offering demonstrates that selective data residency coupled with hyperscaler efficiencies is feasible.
- Cross‑sector collaboration (GSMA, World Bank, CSEP, ministries) is essential to balance trust, efficiency, and innovation while respecting national priorities.
- Audience ideas (data‑embassy rings, multilateral data hubs) illustrate the growing demand for user‑centric, privacy‑by‑design solutions and may shape future policy directions.
Prepared from the verbatim transcript, cleaned and organised for clarity while preserving the nuances of each speaker’s contribution.
See Also: