01.What the regulator has proposed
The Telecom Regulatory Authority of India released draft guidelines on Friday, 30 October 2026, on the responsible use of artificial intelligence in telecom networks and in the provision of telecom services. Stakeholders can send written comments until 27 November 2026. The draft applies to anyone holding a licence under the Telegraph Act, 1885, or an authorisation under the Telecommunications Act, 2023, who uses or plans to use AI.
The asks are concrete. Every operator would have to write an internal AI policy, get it approved by its board or an equivalent body, and file it with TRAI within seven working days of it coming into effect. The policy must be reviewed at least once every financial year. Each AI system would be graded high, medium or low risk based on how much it affects the network, the company's operations and consumers, and its use would need clearance from a designated officer. Companies would also keep an inventory of active, inactive and decommissioned AI systems, and set up an AI Governance and Safety Committee headed by a senior executive designated as the AI governance and safety officer, with members drawn from network operations, service delivery, IT, data management, regulatory and consumer affairs teams.
02.Why telecom got its own rulebook first
AI is already working inside Indian telecom. Airtel runs AI systems that flag spam calls and messages, block malicious links across SMS, OTT apps, email and browsers, and guard against OTP fraud. Jio's in-house platform, JioBrain, is built into network operations for fault prediction, capacity allocation and real-time optimisation. Vodafone Idea uses AI to flag spam and, under Vi Protect, to detect and respond to cyber threats on its core network.
That scale is exactly why the regulator is uneasy. TRAI's draft points to inadvertent biases, unintended model behaviour and adversarial attacks that could affect the resilience of communications infrastructure, which also carries emergency services. The draft is meant to translate the India AI Governance Guidelines released in February 2026 into telecom-specific rules. Those national guidelines rest on seven principles, described as sutras: trust as the foundation, people first, innovation over restraint, fairness and equity, accountability, understandable by design, and safety, resilience and sustainability.
03.The operator answers, even for software it bought
The sharpest line in the draft is about liability. TRAI says the operator will be accountable for the outcome of all AI systems it uses, irrespective of whether they were built in-house, procured from a third party, or a mix of both. To make that workable, companies would have to obtain details of AI embedded in equipment they buy and write audit and accountability clauses into supplier contracts. Before deploying a new system or modifying an existing one, they would need to run due diligence, validation, conformity and impact assessments.
Failures would be graded too, into critical, high, moderate and low severity. Critical and high-severity incidents, including major malfunctions and repeated failures, would have to be reported to TRAI in a prescribed format within a set timeframe. Operators would maintain a central database of AI incidents and a separate record of AI-related harm reported by consumers, along with how each case was resolved. The draft also requires human oversight of AI systems and says data used or generated through a system's lifecycle must be handled under applicable data protection, privacy and cybersecurity rules.
04.What it means for users, students and small firms
For a phone user, the visible change is in redress. The draft says AI-related harms reported by users should be resolved within a reasonable timeframe through a predefined grievance mechanism, and that decisions should be understandable. In practice, that pushes operators to explain why a number was flagged as spam or a message blocked, rather than leaving a subscriber with a silent cut-off and no way to appeal.
For small businesses, the effect runs through procurement. Startups and small firms that sell AI tools, analytics or automation to telecom operators should expect buyers to ask for model documentation, audit rights, incident histories and conformity assessments before signing. That raises the compliance bar for a five-person vendor as much as for a large integrator. For students and young professionals, the draft names roles that barely existed a few years ago: AI governance officers, risk classifiers, incident investigators and auditors who can read a model card and a supplier contract together. These are not research jobs, and they sit in network, legal, compliance and customer operations teams.
05.Soft rules now, a wider law later
It is worth being clear about what the draft is not. These are guidelines under consultation, not binding regulations, and TRAI itself has framed them as light-touch. Satya N. Gupta, a former principal adviser at TRAI, told Mint that they are more guidance than mandate. The draft also notes that other laws continue to apply and that violations would be dealt with under them.
India still has no standalone AI law. It relies on a patchwork: the IT Act, 2000, the Digital Personal Data Protection Act, 2023, consumer protection rules and criminal law. The Ministry of Electronics and Information Technology is expected to bring out a consultation paper on AI regulation by the first week of November, which is expected to form the base for a future law. TRAI had recommended an independent AI regulator back in 2023, and the government has not acted on it. For now, the practical date to watch is 27 November 2026, the last day for comments on the telecom draft.
Frequently asked questions
Does the TRAI draft apply to companies other than Airtel, Jio and Vi?
Yes. It covers anyone holding a licence under the Telegraph Act, 1885, or an authorisation under the Telecommunications Act, 2023, that uses or plans to use AI. That includes registered network providers, not just the big consumer operators.
Are these rules in force already?
No. They are draft guidelines open for stakeholder comments until 27 November 2026. TRAI has described them as light-touch guidance rather than binding regulation, and a separate consultation paper on AI law is expected from MeitY.
What happens if an AI system a telco bought from a vendor fails?
Under the draft, the operator remains accountable for the outcome whether the AI was built in-house or bought. Companies would need audit and accountability clauses in supplier contracts and would have to report critical and high-severity incidents to TRAI.