Government of India Testing Cyber Threats
Linked to AI Model “Mythos”
¨ The Government of India
is testing cyber threats linked to a new AI model called “Mythos” to strengthen
the security of critical digital and financial systems.
¨ The testing includes
Aadhaar identity infrastructure, government login platforms, and
banking-related digital systems.
¨ The initiative involves
technology companies, CERT-In, and several public institutions working
together.
¨ The objective is to
proactively identify vulnerabilities and secure India’s digital infrastructure
against emerging AI-driven cyber risks.
¨
With the rapid growth of
artificial intelligence, cyber security threats are also increasing, prompting
early risk assessment measures.
¨
Anthropic introduced the
next-generation AI model “Claude Mythos Preview” on 7 April 2026.
¨
This model is associated
with detecting and potentially exploiting zero-day vulnerabilities in operating
systems and web browsers.
¨ Zero-day vulnerabilities
are software flaws unknown to developers or for which no security patches are
yet available.
¨ In India, major IT firms
such as Infosys and Tata Consultancy Services are part of the testing
framework.
¨
These companies are
conducting secure simulations using alternative AI models instead of direct
access to Mythos.
¨
CERT-In is actively
overseeing cybersecurity assessments of critical digital infrastructure.
¨
The Reserve Bank of India
held meetings with domestic banks in May 2026 to discuss AI-related cyber
threats.
¨ Under “Project
Glasswing,” Anthropic has provided limited access to around 50 organisations,
including governments and cybersecurity firms.
¨ Reports indicate that the model has identified thousands of potential vulnerabilities, including a significant number of critical ones.
¨ In the United States, institutions such as NIST are also working on AI safety and cybersecurity evaluations.
