India's cyber watchdog is trial-testing open-source artificial intelligence models to proactively spot security vulnerabilities. Developed over six months, this platform aims to shield critical digital infrastructure from increasingly sophisticated automated cyberattacks.
6 months
Development Period
⏳ Time Machine
How today’s news fits into the bigger picture
10 years ago
Back in 2016, cybersecurity meant installing standard firewalls and teaching employees not to click on suspicious email links.
Last year
AI-generated phishing and automated cyberattacks rose, exposing deep vulnerabilities in public sector digital portals.
Last month
Cert-In completed its fifth month of closed-door developmental runs on the new trial platform.
Yesterday
Cyber security audits were static, manual processes that occurred months after a system was deployed.
Today
Cert-In began testing open-source AI models on its newly built automated platform to pre-empt cyber flaws.
What happens next?
By early 2027, developers may need to pass an automated 'AI Safety Rating' before deploying public utility software.
India’s Computer Emergency Response Team (Cert-In) has built a trial platform to put open-source artificial intelligence (AI) models through rigorous cybersecurity testing. Over the last six months, the government watchdog has been quiet about this project, which aims to evaluate how AI models—ranging from text generators to coding assistants—can both locate system vulnerabilities and accidentally leak sensitive data. The timing is critical. As India rapidly digitizes its public services, hackers are increasingly using AI to write malicious code and launch automated attacks. By testing these models in-house, Cert-In wants to establish strict safety benchmarks before these tools are deeply integrated into banking, governance, and utility systems. This move signals a transition from reactive firefighting to automated, predictive defense, turning open-source AI from a potential hazard into a protective shield.
💭 If you're wondering…
Open-source models make their underlying code publicly available, allowing anyone to modify and run them on their own servers. Models like ChatGPT are closed, meaning you must access them through the provider's servers and cannot see or tweak the internal code.
Did this story help?
Official sources
Knowledge Chain — tap a concept