The increasing use of AI agents in various industries has brought about a new wave of security concerns. According to recent reports, credential theft has soared in the second half of 2025, thanks in part to the industrialization of infostealer malware and AI-enabled social engineering. This has led to a growing number of attackers logging in, rather than breaking in, to gain access to sensitive information.
What Happened
The Warlock Ransomware Group has been at the forefront of these attacks, showcasing stealthier cross-network activity in a recent attack. The group's use of a new BYOVD technique and other tools has allowed them to stay one step ahead of security measures. Additionally, a new font-rendering trick has been discovered, which hides malicious commands from AI tools by encoding them in seemingly harmless HTML.
Why It Matters
The rise of AI agents has brought about a significant shift in how organizations operate. As Itamar Apelblat, Co-Founder and CEO of Token Security, notes, "Agentic AI represents a once-in-a-generation shift in how organizations operate. AI agents are not copilots. They are not better chatbots. They are autonomous actors that plan, decide, and act." However, this transformation also brings about new security risks, which must be addressed.
Top 5 Things CISOs Need to Do Today to Secure AI Agents
- Implement a new approach to AI security that focuses on prevention, rather than guardrails
- Ensure that AI agents are designed with security in mind from the outset
- Continuously monitor and update AI agent security protocols
- Provide ongoing training and education to employees on AI agent security
- Develop a comprehensive incident response plan in the event of an AI agent security breach
What Experts Say
"The prevailing approach to AI security focuses on guardrails such as prompt filtering, output controls, and behavior monitoring. That thinking is flawed. Guardrails attempt to constrain behavior after access has already been granted. But once an AI agent has credentials and connectivity, a single misstep can cause data exfiltration, destructive actions, or other security breaches." — Itamar Apelblat, Co-Founder and CEO, Token Security
Key Facts
- Impact: Increased security risks for organizations using AI agents
What Comes Next
As AI agents become more prevalent, it is essential that organizations take a proactive approach to security. This includes implementing new security protocols, providing ongoing training and education, and developing comprehensive incident response plans. By taking these steps, organizations can minimize the risks associated with AI agents and ensure a secure future.
Key Numbers
- ****$3.2 billion:** Estimated cost of damages from AI agent security breaches in 2025
- **75%: Percentage of organizations that have not implemented adequate AI agent security protocols