TRAI AI guidelines for telecom networks were released in draft form on 30 October 2026, setting out the first detailed rulebook for how Indian telecom operators use artificial intelligence. The Telecom Regulatory Authority of India (TRAI) has invited comments from stakeholders by 27 November 2026, and the draft will shape how operators govern AI systems that already run parts of their networks and customer services.
The draft, titled Guidelines on Responsible Use of Artificial Intelligence in Telecommunication Networks and Provision of Telecommunication Services, applies to any entity holding a licence under the Telegraph Act, 1885, or an authorisation under the Telecommunications Act, 2023, to provide telecom services or run telecom networks, where that entity uses or intends to use AI systems. That scope covers Reliance Jio, Bharti Airtel, Vodafone Idea and BSNL, along with smaller licence holders.
Indian operators already use AI across 5G network management, customer service, fraud prevention, network optimisation, predictive maintenance and automated business operations. The TRAI AI guidelines arrive as those uses grow, and as networks move towards more autonomous operation. The regulator warns that incorrect decisions, bias, cybersecurity gaps, service disruption and weak oversight could follow if AI spreads without governance rules.
Board oversight under TRAI AI guidelines
TRAI AI guidelines would require each operator to adopt a board approved AI policy and file it with the regulator within seven working days of it taking effect. Each operator would also set up an AI Governance and Safety Committee with formal board approval, and name a senior executive as AI Governance and Safety Officer to lead it. The aim is to make responsible AI management an enterprise wide duty, not a task left to network or IT teams alone.
Operators would also keep an inventory of every AI system they use, including retired ones, with approval procedures and periodic risk assessments before deployment. Supplier contracts would need enforceable audit and incident clauses before an AI system goes live, because operators would stay accountable for outcomes even when a system comes from an outside vendor.
Risk tiers in TRAI AI guidelines
TRAI AI guidelines sort AI systems into three risk tiers: high, medium and low. The tier depends on how far a system could disrupt connectivity, compromise sensitive data or harm users. High risk systems would face the closest scrutiny, including checks before deployment and stronger human oversight where automated decisions affect critical operations.
The TRAI AI guidelines also set up a four tier incident reporting framework, covering critical, major, moderate and minor events. Operators would report critical or major AI incidents to TRAI within set timeframes, and a central repository would track industry wide AI incident data. The draft points to two Telecom Engineering Centre documents: TEC 57050:2023 for fairness assessment and rating, and TEC 57090:2025 for the incident database schema and taxonomy.
The guidelines are anchored in the India AI Governance Guidelines released by the Centre on 15 February 2026. Those guidelines set out seven principles: trust, people first, innovation over restraint, fairness and equity, accountability, understandable by design, and safety, resilience and sustainability. TRAI’s draft asks sectoral regulators to apply those principles in their own domains, and telecom is the first large sector to get a detailed draft of this kind.
What TRAI AI guidelines mean for operators
TRAI AI guidelines would leave operators accountable for AI outcomes in their networks and services, including systems built by equipment suppliers, software developers, cloud providers and AI companies. That matters for relationships with vendors such as Ericsson, Nokia and Samsung, whose network infrastructure and software now include AI driven automation and optimisation. Operators may need stronger vendor documentation, security testing, explainability and audit support, and may have to write AI governance terms into procurement and vendor agreements.
For Reliance Jio, AI driven network intelligence can support capacity planning, traffic management and service assurance across extensive 4G and 5G infrastructure. Bharti Airtel can apply AI to network optimisation, fraud detection, customer support automation and predictive maintenance. Vodafone Idea can use AI powered network planning as it expands 4G coverage and strengthens 5G capability. BSNL can use AI for network modernisation, fault detection, infrastructure monitoring and customer service improvement. All four would need to check existing and planned AI deployments against the draft’s expectations.
Under the TRAI AI guidelines, operators may need to spend on AI governance platforms, cybersecurity assessments, risk monitoring systems, employee training and independent audits, along with better documentation of models, automated decisions and incident handling. Those costs would land hardest on operators running large networks with many suppliers. The draft remains a proposal. The consultation running to 27 November gives operators, vendors, industry bodies and civil society groups a chance to argue for changes before TRAI decides whether to revise the draft, issue final rules or open a second round of consultation.