TRAI has proposed an AI governance framework for telecom companies, requiring human oversight, board-approved policies, risk classification and reporting of serious incidents involving artificial intelligence.
The Telecom Regulatory Authority of India (TRAI) has proposed a new regulatory framework to promote the responsible use of artificial intelligence across India's telecommunications networks and services.
Under the draft guidelines, telecom operators and other authorised entities would remain accountable for AI-related decisions and outcomes, irrespective of whether their systems are developed in-house or procured from external technology providers.
The proposal aims to strengthen oversight, improve consumer protection and address potential risks associated with the growing use of AI in telecom infrastructure. TRAI has invited comments from stakeholders until November 27. The proposed rules are currently under consultation and have not yet become mandatory.
The draft framework requires telecom companies to establish formal AI governance policies approved by their respective boards.
These policies would need to cover several important areas, including:
Assessment and management of AI-related risks.
Approval procedures before deploying AI systems.
Data governance and protection.
Human supervision of automated processes.
Consumer grievance redressal mechanisms.
Telecom operators would also be expected to establish dedicated governance and safety committees. These bodies would oversee AI deployment and monitor whether systems operate within the prescribed safeguards.
The proposed approach places responsibility on telecom companies to ensure that AI technologies are used safely across their operations.
TRAI has proposed a risk-based classification system for artificial intelligence applications used in the telecom sector.
Under this approach, AI systems would be divided into three categories:
High-risk AI systems: Applications with potentially significant consequences for network operations, financial outcomes or consumers.
Medium-risk AI systems: Systems that present moderate risks and require appropriate monitoring and safeguards.
Low-risk AI systems: Applications with comparatively limited potential impact.
The classification would help determine the level of supervision and safeguards required for different AI applications.
The proposed framework emphasises human supervision, particularly where AI systems influence important decisions or automated processes.
TRAI has cautioned that insufficient oversight could lead to problems such as algorithmic bias, unexpected model behaviour and cyberattacks.
Such failures could affect telecom services and threaten the reliability of critical network infrastructure. Human involvement is therefore intended to help identify problems, assess their impact and take corrective action when necessary.
Another major part of the proposal concerns the reporting and management of incidents involving AI systems.
TRAI has suggested dividing AI-related incidents into four severity levels:
Critical
High
Moderate
Low
Telecom operators would be required to notify TRAI about critical- and high-severity incidents using a prescribed reporting format and within specified timelines.
However, the draft has not yet provided detailed criteria for determining the severity of incidents. It has also not established specific reporting deadlines.
The proposed rules would require telecom companies to maintain centralised databases containing details of AI-related incidents and consumer complaints involving potential harm.
These records would support the investigation of reported problems, help operators implement corrective measures and enable them to track the resolution of complaints.
The proposed mechanism is intended to improve accountability and ensure that AI-related issues are not overlooked.
Artificial intelligence is becoming increasingly important in telecom operations, including customer identification, fraud prevention and network management.
Major Indian operators such as Reliance Jio and Bharti Airtel already use AI-based technologies for customer verification and fraud detection.
Global telecom equipment manufacturers, including Nokia and Ericsson, are also integrating AI capabilities into advanced telecommunications networks.
As these technologies become more deeply embedded in telecom services and infrastructure, regulatory oversight is gaining importance.
The draft framework covers entities holding licences under the Telegraph Act or the Telecommunications Act.
TRAI's proposal seeks to ensure that authorised telecom entities remain responsible for AI-related outcomes, even when they rely on external vendors or third-party technology providers.
This means outsourcing the development or operation of an AI system would not remove an operator's accountability under the proposed framework.
TRAI has opened the draft framework for stakeholder consultation and invited comments until November 27.
The feedback process will allow telecom operators, technology companies, industry representatives and other stakeholders to share their views on the proposed requirements.
The framework is still at the proposal stage. Its provisions may be reviewed or modified after the consultation process before any final regulatory requirements are introduced.
Conclusion
TRAI's proposed AI framework seeks to establish stronger accountability and safety standards as artificial intelligence becomes more widely used across India's telecom sector. Requirements covering board-approved policies, risk classification, human oversight and incident reporting could strengthen consumer protection and network reliability. However, details such as incident severity criteria and reporting deadlines remain to be clarified as the regulator gathers stakeholder feedback.