The emergence of advanced artificial intelligence (AI) systems such as Anthropic’s Claude Mythos is prompting telecom operators to reassess their cybersecurity preparedness, as these tools are increasingly able to uncover software vulnerabilities that traditional systems may miss, senior industry executives said.
Bharti Airtel Chief Technology Officer Randeep Sekhon said the company is in discussions with its suppliers to evaluate potential weaknesses highlighted by such AI-driven assessments.
“We are talking to our suppliers and vendors to understand what vulnerabilities could be exposed and what needs to be fixed,” Sekhon said while speaking to the press. He indicated that operators are working with their technology partners to evaluate and plug such gaps, given that much of the telecom software stack is vendor-owned.
Industry executives said the immediate focus is on assessing the nature of these vulnerabilities and ensuring timely remediation through vendors. Vodafone Idea chief executive Abhijit Kishore said the company is strengthening its network resilience and security frameworks while independently evaluating similar risks. However, he did not confirm any specific external partnerships on the matter.
Despite increased attention on AI-driven vulnerability detection, executives stressed there is no immediate cause for concern. Most of the issues being flagged are understood to be incremental software bugs rather than systemic threats to core telecom infrastructure.
When asked whether the government had reached out to operators on the issue, Sekhon said: “Not yet.” The development comes against the backdrop of Finance Minister Nirmala Sitharaman chairing a high-level meeting with the banking sector on Thursday to review potential risks and strengthen defensive measures.
Telecom networks, which depend heavily on global vendors such as Nokia and Ericsson for core systems, typically rely on these suppliers to roll out patches and security fixes. Operators primarily coordinate implementation to ensure timely updates across networks.
Sekhon also noted that much of what is currently described as AI in telecom operations, including conversational tools replacing traditional search interfaces, represents incremental rather than disruptive change. He said more significant shifts are expected from real-time inferencing, where networks can dynamically respond to changing conditions.
Separately, Airtel acknowledged a rise in memory-related costs impacting server-heavy deployments, though executives said the overall effect on network expenditure remains limited, as not all infrastructure components are memory-intensive.