The Agentic Shift: Why AI-Driven Botnets and Non-Human Identities Define the New Threat Landscape

The Development

The threat landscape has shifted from human-operated campaigns to machine-speed, agentic execution. Recent intelligence confirms the emergence of sophisticated threats like the CARBONATO botnet, which weaponizes exposed Docker services to deploy AI agents directly into compromised environments. These agents, controlled via encrypted channels like Telegram, allow attackers to automate post-exploitation tasks with unprecedented efficiency. This development follows a broader trend of record-breaking ransomware activity, with over 1,000 organizations compromised in August alone, signaling that attackers are successfully integrating AI to scale their extortion operations.

Why It Matters

We are witnessing the erosion of traditional trust models. The integration of AI agents into malware means that attackers no longer need to manually navigate a network; the malware itself can make tactical decisions in real-time. This is compounded by the rise of AI-powered disinformation and deepfakes, which threaten data integrity and non-repudiation. As noted by industry experts, the danger is not just the 'threat of the month' but the systemic distraction caused by these rapid shifts, which often leads organizations to neglect fundamental security hygiene while attackers exploit basic vulnerabilities at machine speed.

Defensive Implications

The shift toward agentic AI requires a fundamental change in how we view 'identity.' In 2026, the perimeter is no longer defined by human users alone. Non-human identities—service accounts, API keys, and autonomous agents—now represent the most significant, yet least protected, attack surface. When an AI agent is deployed inside a network, it can mimic legitimate administrative behavior, making detection via traditional signature-based tools nearly impossible. Defensive strategies must now prioritize behavioral analysis and the strict governance of machine-to-machine communication.

What Leaders Should Do

To maintain resilience in this environment, security leaders must move beyond reactive patching and focus on structural hardening:

• Audit and restrict Non-Human Identities (NHI) with the same rigor applied to privileged human accounts.
• Implement automated, continuous vulnerability discovery to identify exposed services like Docker before they are weaponized.
• Prioritize 'explainable AI' in security tooling to ensure that automated defensive actions can be audited and understood by human analysts.
• Shift focus from chasing individual zero-days to building a robust, federated security architecture that can withstand machine-speed exploitation.

Outlook

The coming months will likely see an increase in 'AI-on-AI' conflict, where defensive systems like the UK’s Cyber Shield are pitted against autonomous offensive agents. As geopolitical tensions remain high, the risk of miscalculation—where error-prone AI systems inadvertently escalate digital or physical conflicts—remains a critical concern. Organizations that fail to secure their automated infrastructure today will find themselves unable to compete with the speed and scale of tomorrow’s adversaries.