Originating from Tel Aviv-Yafo, Israel, and now based in the United States, Tenet Security has secured $6 million in seed funding to advance its mission of mitigating risky AI behaviors in real-time. The startup was founded by Barak Sternberg and Nevo Poran, both former builders of Cisco’s AI Defense research team and alumni of Israel’s elite Unit 8200. The funding round was led by Westly Group, known for its investments in companies like Tesla and SentinelOne.
Revolutionizing AI Security
Tenet Security is introducing a patent-pending technology designed to curb dangerous activities by AI agents, whether they are operating autonomously or have been compromised. These agents can operate invisibly within networks, often evading detection until after they have acted. This innovation is crucial as AI agents become more integral to enterprise operations, presenting new security challenges.
CEO Barak Sternberg emphasizes the transformative potential of AI agents in boosting productivity. However, he also warns of the security risks they pose, as they interact with systems and data in novel ways that traditional security measures cannot comprehend. This creates a new layer of security that necessitates a novel approach to protection.
Addressing Agentjacking and Runtime Threats
The company’s goal is to allow organizations to harness the benefits of AI agents without falling prey to their unpredictable behavior or external manipulations, a threat Tenet terms ‘agentjacking.’ This risk is compounded by the rapid pace at which these agents can operate, often outpacing traditional security responses. Tenet’s solution involves a lightweight runtime sensor that monitors system behavior, network activity, and the reasoning of language models.
By rapidly simulating and predicting an AI agent’s next move, Tenet’s platform can halt harmful actions before they occur. This proactive approach differs from traditional rule-based methods, focusing on preemptive protection rather than reactive measures.
Impact and Future Plans
Research from Tenet’s Threat Labs underscores the severity of agentjacking, identifying thousands of potential vulnerabilities across enterprise environments. Their findings show that malicious actors can exploit agents to bypass conventional security by operating within their authorized permissions, something the Tenet platform effectively counters.
CTO Nevo Poran notes the increasing role of AI agents in attack scenarios, where they can be manipulated to compromise sensitive data or execute unauthorized actions. The challenge lies in real-time monitoring and control of agent behavior at runtime, an area where Tenet’s technology excels.
Early deployments of Tenet’s platform have shown promising results. For example, a major legal-sector enterprise expanded its AI agent usage significantly while thwarting numerous attacks, including a critical cross-site scripting (XSS) attempt. The seed funding will fuel further product development, the expansion of Tenet Threat Labs, and the company’s North American market operations, as well as broaden their coverage of new AI agent frameworks.
