India Tests Banks and Government Systems as Anthropic Mythos AI Raises Cybersecurity Alarms

India Mobilizes to Test Critical Infrastructure Against AI Threats

India has initiated comprehensive security assessments of its most sensitive digital systems in response to growing concerns about Anthropic Mythos, an advanced AI model with unprecedented capability to identify software vulnerabilities 1

1

. Essential services providers including banks, telecom operators, and power utilities have begun stress-testing critical software systems to prepare for the model's broader launch expected within six to seven months 1

1

. The urgency stems from Mythos identifying 23,019 vulnerabilities in just one month, while only 97 have been patched as human cybersecurity professionals require an average of 14 days to fix each bug 1

1

.

Major Tech Firms Lead Security Reviews

Indian technology giants Infosys and TCS are among companies conducting vulnerability tests in secure environments, according to officials familiar with the matter 2

2

. Infosys is specifically focused on devising patches for its widely used Finacle banking software, which provides services to financial institutions globally 2

2

. Since these companies lack direct access to Anthropic Mythos, they are using Anthropic's Claude Opus 4.7 AI software to identify and patch vulnerabilities 2

2

. Companies are also testing their public-facing assets using existing advanced AI models such as Opus 4.7 and GPT 5.5, while asking suppliers to identify possible gaps 1

1

.

Government Agencies Examine National Digital Infrastructure

CERT-In, India's state-run cybersecurity agency, is conducting tests of key digital infrastructure including the Aadhaar ID program and government login systems 2

2

. The Data Security Council of India (DSCI), an industry think-tank under Nasscom, has launched a sandbox environment for organizations to evaluate generative AI models against potential security vulnerabilities and data privacy risks 1

1

. This proactive governance approach reflects the scale of potential exposure across critical national systems including banking, telecom, healthcare, cloud infrastructure, and energy networks 1

1

.

Banking Sector Confronts Unprecedented AI-Driven Threats

The Reserve Bank of India has held multiple meetings with banks to apprise them of potential threats from Mythos 2

2

. Finance Minister Nirmala Sitharaman asked banks to step up vigilance to secure IT systems, safeguard customer data, and protect financial resources 2

2

. A CISO at a large private sector bank explained that the scale and velocity at which systems like Mythos can detect vulnerabilities fundamentally change the threat landscape, compressing what previously took weeks or months into hours 1

1

. The widening gap between detection and remediation poses the real concern, as banks still rely heavily on manual patch management cycles, legacy infrastructure dependencies, and fragmented vendor ecosystems 1

1

.

Talent Gap Compounds Cybersecurity Challenges

Indian banks are operating in an environment with a significant talent gap in cybersecurity expertise 1

1

. A CISO with a state-run lender noted that security teams were never designed to handle tens of thousands of vulnerability alerts at machine speed, creating risks of alert fatigue and prioritization failure 1

1

. Critical vulnerabilities in internet banking, telecom integrations, or third-party fintech systems could remain unpatched because security teams are overwhelmed 1

1

. Experts warn that cybercriminals might exploit vulnerabilities faster than organizations can fix them, especially as Mythos-capability models from DeepSeek or OpenAI are expected to hit the market in three to six months 1

1

.

India Seeks Controlled Access to Mythos

New Delhi is seeking access to Anthropic Mythos in a secure setup preferably based within India, with talks led by the Ministry of External Affairs engaging with US government officials 2

2

. Anthropic has told Indian authorities that it's for the US government to decide if they would share the advanced model with any company or country 2

2

. Currently, Anthropic has limited access to Mythos to select companies including Apple and JPMorgan Chase under Project Glasswing 2

2

. India sees closer cooperation with the US as key to protecting sensitive infrastructure and government networks while reducing reliance on rivals such as China 2

2

.

Organizations Strengthen Cyber Hygiene

Vinayak Godse, chief executive of the Data Security Council of India (DSCI), stated that organizations are actively strengthening their cyber hygiene to prepare for AI-driven threats by reducing attack surfaces, adopting micro-segmentation, and improving identity and authentication systems 1

1

. Srikanth Velamakanni, chief executive at Fractal Analytics and chairperson of Nasscom, emphasized that systems like Mythos demonstrate that releasing highly capable models without stringent safety frameworks can pose serious national security threats, calling for proactive governance rather than simply attempting to contain AI development 1

1

. Infosys CEO Salil Parekh told analysts that Mythos was "exposing more vulnerabilities than one thought possible previously," while also noting it could lead to new opportunities to help clients address vulnerabilities 2

2

.