India’s telecom regulator has proposed new governance and incident-reporting duties for applicable authorized telecom entities that use artificial intelligence, seeking to address risks to services and network infrastructure.
The Telecom Regulatory Authority of India (TRAI) issued the draft Oct. 30 and is accepting comments through Nov. 27. It covers licensed telecom service providers and entities authorized to establish, operate, maintain or expand networks when they use or plan to use AI in networks or services.
Each covered entity would need an internal AI policy approved by its board or an equivalent governing body. The policy would classify systems as high, medium or low risk, weighing their significance to telecom networks and service provision, financial and operational outcomes, and potential effects on consumers. Entities would review the policy at least once each financial year.
A designated officer would have to approve each AI system before it is used. The draft also calls for validation, conformity and impact assessments before deployment or modification, along with human oversight and secure, lawful data handling. Entities would remain accountable for system outcomes whether the technology was developed internally, obtained from a vendor or built through both approaches.
TRAI would require each entity to establish an AI Governance and Safety Committee led by a senior executive. All AI incidents would need to be recorded and investigated, while critical- and high-severity cases would be reported to TRAI. The proposal also calls for one central database for AI incidents and a separate central database for reports of AI-related harm to telecom users and their resolution.
The regulator says AI is increasingly being integrated into telecom networks and services, and presents the proposal as a sector framework for consistent implementation of India’s broader AI governance guidelines. It says AI may improve telecom efficiency, while biased results, shifts in model behavior and deliberate attacks could weaken network resilience, especially in critical components.
Nokia said it introduced, with NVIDIA, a platform for telecom operators to build AI agents across network operations.
The draft assigns policy, oversight, record-keeping and reporting work to covered entities but gives no estimate of compliance costs. Stakeholders can send comments to TRAI Advisor Akhilesh Kumar Trivedi at advmn@trai.gov.in by Nov. 27. The draft sets no date for a final decision.
