The Indian Banks' Association has begun discussions with banks to assess potential risks from Anthropic's advanced AI model, Mythos, amid concerns over its ability to detect software vulnerabilities.
The industry body held a meeting with senior bank officials in Mumbai on April 29 to examine possible safeguards against threats linked to the model, sources told NDTV Profit. Officials from CERT-In also met bankers on the issue. Lenders are considering appointing global technology firms such as Microsoft and IBM to prepare for risks associated with Mythos.
The discussions reflect growing concern within the banking system over the model's ability to identify weaknesses in widely used software, which could expose financial institutions to cyber risks.
Access and oversight
Banks are looking to work with third-party technology providers as the finance ministry is not in favour of giving lenders direct access to Mythos, according to sources. Earlier this month, Anthropic launched Project Glasswing, a restricted programme aimed at securing critical software infrastructure using Mythos Preview. The initiative includes selected US banks and regulators.
Last week, Finance Minister Nirmala Sitharaman met senior bank executives in New Delhi to review the issue. She said the IBA has been tasked with assessing risks from the model and directed banks to strengthen internal IT systems to manage cyber threats. The Reserve Bank of India is also engaging with global regulators on developments related to Mythos.
Capability concerns
Mythos is designed to identify vulnerabilities in operating software. Bankers are concerned about its ability to detect bugs at scale and fear they may not be able to address such vulnerabilities in real time.
ALSO READ: CERT-In Issues 'High-Severity Alert' Amid Mythos Jitters; MSMEs Urged To Shield Against AI Risks
Essential Business Intelligence, Continuous LIVE TV, Sharp Market Insights, Practical Personal Finance Advice and Latest Stories — On NDTV Profit.
