Key Details
TRAI’s draft Guidelines on Responsible Use of Artificial Intelligence in Telecommunication Networks and Provision of Telecommunication Services translate India’s broader AI-governance principles into proposed responsibilities for telecom networks and services.
Area | What the Draft Proposes |
|---|---|
Who is covered | Entities licensed or authorised to provide telecom services or operate telecom networks if they use or intend to use AI |
Internal oversight | A governing-body-approved AI policy and a committee led by a senior AI Governance and Safety Officer |
Before use | Risk classification, designated approval, supplier checks, and validation and impact assessment of new or modified AI systems |
During use | An inventory of AI systems, appropriate human oversight, incident records and a process for resolving user-reported harm |
Serious incidents | Reporting of critical- and high-severity AI incidents to TRAI; the form and timeframe would be prescribed |
Accountability | The authorised entity remains accountable for AI it uses even when a third party supplies it |
AI Used in Telecom Networks Would Face Checks Before and After Deployment
AI can help operate networks and deliver services, but an error in a consequential system may affect connectivity or users at scale. TRAI therefore proposes that each covered entity classify its AI systems as high, medium or low riskaccording to their significance for network operations, services and consumers. New systems — and changes to existing one — would undergo assessment before deployment, followed by human oversight and periodic review.
The draft also requires entities to keep an inventory of active, inactive and retired AI systems. This would make it possible to identify which systems are in use, what they do and who approved them, rather than treating AI as an untracked feature within equipment or software.
Operators Would Retain Responsibility for Supplier-Built AI
Telecom entities often acquire systems from outside vendors. The draft would require checks on suppliers and their AI systems, disclosures when telecom equipment contains AI, and contracts covering matters such as audit rights, incidents and accountability. Crucially, outsourcing would not transfer responsibility away from the authorised entity using the system.
For incidents, entities would maintain central records, investigate failures and report the most serious cases to TRAI. They would also record and resolve AI-related harm reported by users. TRAI suggests relevant technical standards for assessing fairness and recording incidents, but the draft does not make every example standard universally mandatory.
Consultation Will Determine the Final Requirements
TRAI issued the draft to apply the India AI Governance Guidelines, released in February 2026, to a sector on which essential services and emergency response depend. Stakeholders can comment by 27 November 2026. The proposed duties, including the details of serious-incident reporting, may change before guidelines are finalised.
Policy Relevance
The draft’s principal shift is from broad AI principles to identifiable responsibilities within telecom entities: approval before use, continuing supervision, records of failures and accountability for vendor-supplied systems. Consultation should clarify how risk categories will work across very different AI uses, what constitutes a reportable incident and what information TRAI will need without duplicating existing legal or licence obligations.
Relevant Question for Policy Stakeholders: How should TRAI distinguish high-risk AI uses in critical network operations from lower-risk applications, so that oversight is proportionate without leaving consequential failures unreported?
Follow the Full Draft Consultation and Press Release Here: TRAI’s Draft Guidelines on Responsible Use of AI in Telecom | TRAI’s Consultation Announcement