Gujarat-based company develops ‘AI Action Firewall’ to make artificial intelligence systems safer
Gujarat-based company develops ‘AI Action Firewall’ to make artificial intelligence systems safer
Publish Date: 2026-03-21 09:12:00
Source Domain: m.economictimes.com
- The Gujarat-based neuro-engineering company NeuroPause Lab Limited has developed an “AI Action Firewall” to enhance the safety of AI systems.
- The firewall works as a policy-based safety layer to ensure every AI action is authorised, monitored, and recorded.
- It draws parallels with traditional firewalls but is specifically designed to control AI-driven actions such as emails, code execution, and database access.
- NeuroPause Lab Limited’s CEO, Saurabh Patel, highlighted the company’s patent application for this technology last year.
- Each AI action will be evaluated against predefined policies, which will classify actions as “allow,” “review,” or “block.”
- Higher-risk operations will require human approval, while unauthorised actions will be stopped to prevent misuse.
- All decisions made by the firewall will be logged for transparency, accountability, and regulatory compliance.
- Patel emphasized that existing cybersecurity systems are unsuitable for managing AI agents, requiring a dedicated control layer for AI actions.
