Cato Networks Launches $30,000 AI Security Challenge
Translated & summarized from Ice by baba
Cato Networks, led by Shlomo Kramer, has launched a challenge offering $30,000 in prizes for successfully outsmarting its new AI security solution, AgentSense. The challenge aims to test and improve the system's ability to prevent AI agents from performing malicious actions. AgentSense monitors AI agents in real-time to ensure their actions align with user intent, addressing a critical security gap as organizations increasingly rely on AI for business operations.
The story in 5 lines · by baba
- Cato Networks is offering $30,000 in prizes for successfully outsmarting its new AI security agent.
- The challenge tests AgentSense, a new patented solution designed to monitor and control autonomous AI agents.
- AgentSense aims to prevent AI agents from deviating from their intended tasks due to malicious instructions.
- The company will use challenge findings to improve AgentSense's defensive capabilities against AI threats.
- CEO Shlomo Kramer emphasized the critical need for AI agent oversight in business operations.
Cybersecurity firm Cato Networks, founded and led by Shlomo Kramer, has announced a unique challenge for the information security community, offering a total prize of $30,000 to individuals who can successfully outsmart an artificial intelligence (AI) agent. The first three qualifying participants to complete the task will each receive $10,000.
The challenge coincides with the company's unveiling of AgentSense, a new, patented security solution designed to continuously monitor autonomous AI agents in real-time. AgentSense aims to ensure that AI agents' decisions, actions, and outputs align with user intent. Participants in the challenge, accessible via the company's website, are invited to attempt to compromise an AI agent's workflow protected by AgentSense. The objective is to embed malicious instructions within content exposed to the AI agent, causing it to deviate from its designated task.
Cato Networks stated that findings from the challenge will be used to refine and enhance AgentSense's defensive capabilities. The company explained that even AI agents operating with legitimate permissions can perform risky actions beyond their defined tasks. A core vulnerability lies in their difficulty distinguishing between valid instructions and external content encountered during operation. Attackers can exploit this by hiding malicious commands in emails, web pages, documents, or tool responses.
When an AI agent has access to sensitive data and the ability to execute actions independently, it creates what Cato Networks calls the "deadly trinity," a significant business threat that could lead to data breaches, system damage, and operational disruptions. Unlike solutions relying on known attack pattern recognition, AgentSense analyzes in real-time how untrusted content influences an agent's decisions and actions, offering protection against unknown vulnerabilities, including zero-days.
Shlomo Kramer, CEO and co-founder of Cato Networks, commented on the launch and challenge, stating, "In an era where organizations entrust AI agents with decisions impacting revenue, customer relations, and business continuity, controlling agent activity is as crucial as financial oversight. Granting AI agents greater autonomy necessitates security mechanisms to thwart harmful actions. Our vision is for AgentSense to give our customers the confidence to rely on AI agents for critical business processes, without losing control and overcoming the inherent challenges."
Cato Networks provides a unified cloud platform integrating networking, security, and access management, enabling organizations to replace legacy infrastructure, understand user-application-data-AI interaction contexts, and adopt AI tools while maintaining operational resilience and high security.
The same event, reported separately by each outlet. Open a few to compare what different newsrooms emphasize — and what they leave out.
Other 2
Not the same event — other stories that share this one’s people, places, or theme: background, reactions, and follow-ups.
