The Telecom Regulatory Authority of India (TRAI) has proposed a new framework requiring telecom operators to classify artificial intelligence systems by risk, establish internal AI governance policies and report serious AI-related incidents to the regulator.
The draft Guidelines on Responsible Use of Artificial Intelligence in Telecommunication Networks and Provision of Telecommunication Services were released on October 30 for public consultation.
The proposals seek to make telecom companies accountable for the outcomes of AI systems deployed in their networks and services, regardless of whether the technology is developed internally or purchased from external vendors.
TRAI has invited written comments from stakeholders until November 27, 2026.
Telecom Operators May Need Board-Approved AI Policies
Under the proposed framework, telecom operators and other authorised entities using AI in their networks or services will have to formulate an internal AI policy approved by their board of directors or equivalent governing body.
The policy must classify every AI system as high-risk, medium-risk or low-risk.
The classification will depend on the system’s importance to telecom operations, its impact on financial and operational outcomes, and its potential consequences for consumers.
The draft also proposes that AI systems cannot be deployed without approval from an officer designated by the organisation.
Companies will have to maintain an updated inventory of all active, inactive and decommissioned AI systems, recording their sources, purposes and risk classifications.
The policy must be reviewed at least once every financial year. Its initial version and subsequent amendments must also be filed with TRAI within seven working days of coming into effect.
Serious AI Failures May Have To Be Reported To TRAI
One of the important provisions concerns incidents involving AI systems.
The draft proposes four categories of AI incidents: critical, high, moderate and low severity.
Telecom operators will have to maintain a central database recording these incidents, investigate their causes and implement measures to reduce the resulting harm.
Critical and high-severity incidents must be reported to TRAI in a prescribed format and within a prescribed timeframe.
This could cover material malfunctions, recurring system failures and the corrective measures taken by operators.
However, the draft does not yet clearly define what constitutes a critical or high-severity incident or specify the reporting deadlines. These details are left to be prescribed.
The proposal is intended to give the regulator greater visibility into AI-related failures affecting telecom networks and services.
Companies Remain Accountable For Third-Party AI
The proposed framework also addresses the growing dependence of telecom operators on external technology providers.
Telecom companies increasingly purchase AI-enabled software and equipment rather than developing every system themselves.
TRAI’s draft states that an authorised entity will remain accountable for the outcomes of all AI systems it uses, irrespective of whether they are developed internally, procured from a third party or jointly developed.
Before acquiring an external AI system, operators will have to examine the credibility of the supplier, the system’s capabilities and limitations, and the quality of its training data.
Supplier agreements will also need provisions covering regulatory compliance, audit and inspection rights, incident management, risk allocation and accountability.
This means operators may not be able to shift responsibility to a technology vendor simply because the AI system was developed outside their organisation.
Human Oversight And Consumer Protection Proposed
The draft requires appropriate human oversight of deployed AI systems as part of risk management.
Before introducing a new AI system or modifying an existing one, operators will have to undertake validation, conformity assessments and impact assessments.
These assessments must consider accuracy, robustness, explainability, fairness and the system’s ability to withstand changes in its operating environment.
The guidelines also propose central databases for recording consumer complaints involving AI-related harm.
Operators will have to establish grievance redressal mechanisms and resolve reported harm within a reasonable timeframe.
For telecom users, this provision could become particularly relevant as AI increasingly influences services such as spam detection, customer identification, fraud prevention and network management.
If an AI-driven system produces an incorrect or harmful outcome, the proposed framework places responsibility on the service provider to investigate and address the complaint.
Draft Does Not Introduce A Separate Penalty Framework
Although the proposed guidelines repeatedly use mandatory language, they do not establish a separate penalty schedule for non-compliance.
Instead, the draft states that violations involving AI systems that also breach existing laws, regulations, licence conditions or other binding requirements will be dealt with under the relevant provisions.
The distinction matters because the document is currently a consultation draft and should not be treated as an already implemented regulatory framework.
TRAI has issued the proposals under its existing regulatory mandate, rather than introducing a separate AI law for the telecommunications sector.
The regulator is seeking to apply the broader India AI Governance Guidelines, released in February 2026, to the specific risks associated with telecom infrastructure.
Stakeholders Have Until November 27 To Respond
The draft applies to entities holding relevant telecom licences or authorisations that use or intend to use AI in telecom networks or services.
TRAI has invited written comments from industry stakeholders by November 27, 2026.
Submissions may be sent electronically to Akhilesh Kumar Trivedi, Advisor (Networks, Spectrum and Licensing), at advmn@trai.gov.in.
The consultation will help determine how the proposed requirements are refined before the framework moves towards finalisation.
As AI becomes increasingly embedded in India’s telecommunications infrastructure, the proposals seek to establish a clearer line of responsibility between technology providers, telecom operators and the regulator.
For consumers, the eventual framework could influence how AI-driven telecom services handle personal data, detect fraud, manage network operations and respond when automated systems cause harm.
