India's nodal cybersecurity agency has issued a high-severity advisory warning that frontier AI systems can now autonomously discover vulnerabilities, orchestrate multi-stage attacks, and execute social engineering at machine speed. Here's what you need to know — and do.
The Indian Computer Emergency Response Team (CERT-In) has released one of its most consequential advisories to date — and this time, the threat isn't a single CVE or a rogue malware strain. It's the convergence of frontier Artificial Intelligence with offensive cyber operations.
Advisory CIAD-2026-0020 signals a tipping point: AI systems have matured to a level where they can autonomously find, exploit, and chain vulnerabilities at a speed and scale that previously demanded entire teams of expert attackers. For CISOs, GRC professionals, and business leaders — this is not a future risk to prepare for. It is a present reality to respond to.
CERT-In's advisory draws on emerging research and evaluations of next-generation AI models, pointing to a category shift in attacker capability. The key capabilities that frontier AI now brings to adversaries include:
⚠️ CERT-In Risk Summary: These capabilities lower the barrier to entry for malicious actors dramatically. Organisations that once felt secure because attackers "wouldn't bother" are now equally viable targets for fully automated, low-cost campaigns.
The advisory is direct about potential consequences. A successful AI-driven attack on your environment could result in:
The advisory provides a comprehensive set of actionable recommendations. Here is a structured summary of the six pillars CERT-In has prescribed for enterprises:
Elevate security monitoring frequency. Remove unnecessary internet-facing services. Tune detection tools for AI-driven attack signatures — unusually fast scanning, anomalous access patterns, unfamiliar scripts. Treat every newly disclosed critical vulnerability as exploitable within hours, not weeks. Enable DDoS protection on all internet-facing assets.
Enforce MFA across all internet-facing services, remote access gateways, and cloud consoles. Mandate hardware-based identity for access to sensitive systems — stolen credentials alone must never grant entry. Implement micro-segmentation to prevent lateral movement. Review and harden legacy VPN infrastructure.
Target critical patch deployment within 24 hours of release for internet-facing systems. Automate patch intake and triage. Maintain current IT asset inventories. Evaluate open-source dependencies with tools like OpenSSF Scorecard. Extend patching SLAs to vendors and the broader supply chain. Track Software, AI, and Hardware Bill of Materials (SBOM/AIBOM/HBOM).
Enforce strong password policies, remove default credentials, and disable unused services and ports. Apply the 3-2-1 backup rule with regular restoration testing. Deploy updated endpoint protection. Encrypt data at rest and in transit. Monitor and restrict outbound traffic to AI service endpoints to prevent unsanctioned data sharing with external AI tools.
Train security teams on AI-augmented attacker tactics. Run realistic phishing simulations that include AI-generated voice, video, and text lures. Designate AI Security Champions in each business unit. Conduct external AI red-teaming — unauthenticated, no source access — to simulate real attacker conditions.
Update IR and Cyber Crisis Management plans for accelerated, multi-front exploitation scenarios. Pre-arrange forensics and IR retainer agreements for immediate activation. Conduct tabletop exercises that simulate five simultaneous incidents, modelling AI-driven scenarios. Strengthen BCP/DR through tested recovery procedures. Report suspicious activity and preserve logs per CERT-In Directions 2022.
The advisory dedicates specific guidance to Micro, Small, and Medium Enterprises — recognising that resource constraints are real but not an excuse. CERT-In recommends MSMEs focus on: enabling automatic OS and application updates, adopting MFA, using managed security services for patching and monitoring, avoiding unverified AI tools in production, and building a structured cyber incident response plan. Regular employee training on AI-generated scams is particularly emphasised.
At CyRAACS, we believe this advisory reinforces something we've long advocated: Governance, Risk, and Compliance is not a checkbox exercise — it is operational infrastructure for cyber resilience.
The six pillars CERT-In recommends map directly to the capabilities our COMPASS GRC platform is built to enable. From continuous vulnerability tracking and vendor risk management, to policy lifecycle governance and incident response workflow automation — COMPASS is designed to give organisations the structured visibility they need to respond to exactly the kind of accelerated, AI-driven threat environment CERT-In is now warning about.
The question for every organisation is no longer if they will be targeted by an AI-augmented attack — it is whether their GRC posture is mature enough to detect it, contain it, and recover from it faster than the attacker can adapt.
See how COMPASS by CyRAACS helps you operationalise CERT-In's recommendations — from patch governance to incident response readiness.
Request a COMPASS Demo