India's telecom regulator has published draft guidelines for the responsible use of artificial intelligence in telecom networks and services. The Telecom Regulatory Authority of India (TRAI) released the draft on October 30, 2026, and stakeholders have until November 27, 2026 to submit comments before the regulator finalises the framework.
The draft, titled "Guidelines on Responsible Use of Artificial Intelligence (AI) in Telecommunication Networks and Provision of Telecommunication Services", applies to authorised entities that use, or plan to use, AI in telecom networks or in providing telecom services. It follows the government's India AI Governance Guidelines, released in February 2026.
What the draft requires
At the centre of the proposal is a board-approved internal AI policy. Each operator would have to set out how its AI systems are classified, approved, monitored and governed, and file the policy with TRAI within seven working days of it taking effect, reviewing it at least once every financial year.
Operators would also have to form an AI Governance and Safety Committee, approved by the board and headed by a senior executive designated as the AI Governance and Safety Officer. The committee would draw members from network operations, service delivery, IT, data, regulatory and consumer affairs, and would review AI performance and incident records at least annually, with the power to require a system to be modified or decommissioned.
Every AI system would be sorted into a risk tier — high, medium or low — based on its materiality to network operations, service delivery, financial and operational outcomes, and its potential impact on consumers. No system could be used without sign-off from a designated officer. Operators would also have to keep an inventory of active, inactive and decommissioned AI systems, carry out due diligence before procuring systems from third-party suppliers, and complete validation and impact assessments before deploying or modifying a system.
Accountability follows the operator, not the vendor
The sharpest provision in the draft is its accountability rule: the authorised telecom entity is accountable for the outcome of every AI system it uses, irrespective of whether the system was developed internally, procured from a third party, or assembled from both. A supplier contract may divide costs or indemnities, but not accountability.
Human oversight of AI systems would be mandatory. Every AI-related incident would have to be recorded, investigated and analysed, and classified into one of four severity bands — critical, high, moderate and low — with critical and high-severity incidents reported to TRAI.
The provision matters because operators already deploy AI widely: for network management, traffic optimisation, predictive maintenance and customer service. When an automated system misroutes traffic, mishandles a subscriber's data or makes a faulty commercial decision, the draft's answer is that the operator, not the vendor that sold it the model, must answer for it.
Guidance, not yet law — and why it matters
It is important to be clear about the draft's status: this is a proposal open for consultation, not a final regulation. Former TRAI principal adviser Satya N. Gupta has described the guidelines as more in the nature of guidance than enforceable mandates, serving as a code of conduct under light-touch regulation.
Still, the direction is unmistakable. By framing AI governance as a boardroom responsibility rather than a purely technical one, TRAI is signalling that operators must be able to explain, audit and stand behind the automated systems running their networks. The submissions received by November 27 will shape whether that signal hardens into binding rules — and how quickly operators must reorganise to comply.
