India proposes mandatory reporting of serious telecom AI failures
India’s telecom regulator has proposed making operators report serious AI incidents and answer for systems bought from vendors. A board-approved policy would classify risks and set human oversight arrangements. The consultation runs until November 27, with reporting procedures and supplier accountability among the proposed duties for networks already using AI.
Artificial Intelligence··Midday
Operators would report serious failures
The Telecom Regulatory Authority of India, known as TRAI, has opened consultation on rules for artificial intelligence in telecom networks and services. Operators would report critical and high-severity incidents, including major malfunctions and repeated failures. The proposal covers licensed or authorized businesses that already use AI or plan to introduce it. Written comments are due on November 27. The obligations remain proposals pending the consultation.[1], [2]
Boards would approve AI policies
A company’s board would approve its internal AI policy and submit it to the regulator within seven working days. The policy would identify who can authorize use, classify systems as high, medium or low risk, and set arrangements for human oversight. It would also cover data handling, incident grading and complaints from users. Annual review is proposed, alongside a governance and safety committee drawing members from network, technology, regulatory and consumer-affairs teams.[1]
Vendor systems remain the operator’s responsibility
The draft makes operators accountable for AI acquired from suppliers as well as systems developed internally. Companies would obtain information about AI embedded in purchased equipment and include audit and accountability provisions in supplier contracts. Such systems already perform network tasks: Jio uses its JioBrain platform for fault prediction and capacity allocation, while Airtel applies AI to spam detection and malicious-link blocking. Existing Indian laws would continue to apply under the proposed guidelines.[1]
