News · eGov
TRAI has released draft guidelines on responsible AI use in telecom networks, proposing risk classification, incident reporting, board-approved policies and governance committees, with stakeholder comments due by 27 November 2026.
✦ AI SUMMARY 30-second read
- Draft guidelines propose that telecom operators classify AI systems by risk level — high, medium or low — and maintain a live inventory of all active, inactive and decommissioned systems.
- Operators would be required to form an AI Governance and Safety Committee, led by a senior executive designated as AI Governance and Safety Officer, with at least annual reviews of deployed AI systems.
- Critical and high-severity AI incidents must be reported to TRAI in a prescribed format; written stakeholder comments on the draft are accepted until 27 November 2026.
Why it matters: Telecom networks underpin India's digital economy and critical infrastructure, making accountable AI governance in the sector a significant public interest concern.
This summary was generated by AI. Editors may review and update it.
The Telecom Regulatory Authority of India (TRAI) has issued a draft framework titled Guidelines on Responsible Use of Artificial Intelligence (AI) in Telecommunication Networks and Provision of Telecommunication Services. The document sets out proposed rules covering AI risk classification, incident reporting, board-approved governance policies and consumer grievance mechanisms for telecom operators.
Stakeholders have until 27 November 2026 to submit written comments on the draft. The proposals are consultative and are not to be read as final regulatory requirements.
The initiative follows the Government of India's AI Governance Guidelines, which were released in February 2026 and rest on seven principles: trust, people-first development, fairness, accountability, safety and resilience, and transparency of design. TRAI's draft seeks to apply that national framework specifically to the telecommunications sector.
What was announced
The draft proposals apply to AI systems that telecom operators have either developed in-house or procured from third parties. Operators would be required to maintain a register of all such systems — active, inactive and decommissioned — and no system could be used unless it had been added to the inventory.
Risk classification would categorise each AI system as high, medium or low, based on factors including the materiality of the system and its potential impact on the operator's operations, finances and consumers. Operators would need to evaluate AI systems for accuracy, robustness, explainability, fairness, resilience to model change and the effectiveness of human oversight before deployment or modification.
The framework also proposes a structured incident management process. Incidents would be graded at four severity levels — critical, high, moderate and low — and operators would maintain a master incident database. Critical and high-severity incidents would have to be reported to TRAI in a prescribed format within a stipulated timeframe.
Governance structure proposed
Each covered operator would be expected to establish an AI Governance and Safety Committee (AGSC) with board approval. The committee would draw members from network operations, information technology, regulatory affairs and consumer affairs functions.
A senior executive designated as AI Governance and Safety Officer (AGSO) would lead the committee. The AGSC would review AI system deployments at least annually, examine incident records, assess emerging risks and, where necessary, recommend changes to or decommissioning of a system.
A board-approved AI policy would also be required, covering risk classification, system approval, incident response, human oversight, data governance and consumer grievance redressal. Operators would need to submit the policy to TRAI within 7 working days of its implementation and conduct periodic reviews.
Third-party suppliers and consumer protection
The draft extends its scope to AI systems sourced from external vendors. Operators would be expected to carry out supplier due diligence and embed compliance, audit, incident management and accountability obligations into supplier agreements.
On consumer protection, the proposals address lawful and secure data processing, privacy and confidentiality safeguards, and redress for AI-related harms. Operators would be required to maintain records of consumer complaints and their resolution.
What happens next
TRAI has invited written feedback on the draft until 27 November 2026 and has directed stakeholders to the regulator's official website for the draft document and submission guidelines.
The proposals will evolve through the consultative process before any final guidelines are issued. The precise obligations for telecom operators will become clearer once TRAI concludes the consultation and finalises the framework.
Elets will update this story as the regulatory process advances.