Policy experts argue that India should create a formal mechanism giving government agencies, technical institutions, and independent researchers early access to frontier AI models in controlled environments for safety and red-team testing. Findings could be shared confidentially with developers so vulnerabilities can be addressed before extensive deployment.