AI trust, safety, and security company Alice, formerly known as ActiveFence, recently announced a significant funding round, raising $140 million. This latest capital injection brings their total funding to $280 million, underscoring the increasing industry focus on securing artificial intelligence systems against emerging vulnerabilities and adversarial attacks. The investment, led by Apax Digital Funds with participation from several other firms, will be channeled into advancing Alice’s AI platform, expanding its proprietary intelligence dataset, and scaling its market presence, according to SecurityWeek.
This funding arrives at a critical juncture as organizations rapidly deploy AI, particularly generative AI models, which introduce new attack surfaces and manipulation vectors. Alice’s mission addresses the growing imperative to ensure these powerful technologies operate safely and within defined ethical and business parameters.
Technical Approach to Securing AI Systems
Alice collaborates with AI developers to proactively stress-test foundational models before their public release. Their research division simulates malicious inputs and complex agentic tasks, specifically designed to expose erratic behaviors and harden models against common exploits like prompt injection and jailbreak attempts. This pre-deployment validation is crucial for identifying and remediating weaknesses before they can be exploited in the wild.
Once AI models transition into production environments, Alice provides continuous red-teaming services and dynamic runtime guardrails. These guardrails are engineered to align AI operations with an organization’s specific business rules and internal guidelines. Companies can configure custom policies, simulate potential breaches, and monitor system traffic in real time to maintain control and visibility over their AI deployments. For security professionals concerned with detecting prompt injection attempts and other adversarial inputs, understanding such continuous monitoring and red-teaming capabilities is vital.
A core component of Alice’s defense strategy is its proprietary database, Rabbit Hole. This extensive repository has been meticulously built over nearly a decade, tracking digital fraud, ideological extremism, and manipulation campaigns across various platforms. By leveraging this vast archive of harmful content, the platform is equipped to detect and intercept hostile actions targeting generative AI systems. Furthermore, Alice operates a dedicated research facility with over 150 specialists who work alongside major model developers to anticipate failures, analyze potential manipulation vectors, and proactively close security gaps in machine learning architectures.
The Growing Imperative of AI Model Defenses
The rapid adoption of generative AI systems across industries presents both unprecedented opportunities and significant security challenges. As Alice CEO and co-founder Noam Schwartz notes, “There are infinite ways to break an AI, and you can’t defend against something you’ve never seen.” This highlights the complex, evolving nature of AI security, where traditional cybersecurity paradigms often fall short. The funding will help close the gap between the rapid democratization of AI capabilities and the comparatively slower development of comprehensive defenses. Organizations are increasingly looking for effective strategies for mitigating generative AI vulnerabilities as these technologies move from experimental phases to core business functions.
Actionable Recommendations for Implementing Enterprise AI Guardrails
For security professionals tasked with safeguarding AI initiatives, several priorities emerge:
- Prioritize Pre-deployment Testing: Engage in comprehensive stress-testing and red-teaming of AI models before they are deployed to identify and address vulnerabilities like prompt injection.
- Implement Runtime Monitoring: Establish continuous monitoring and dynamic guardrails for AI systems in production. This allows for real-time detection and mitigation of adversarial attacks and policy violations.
- Define Clear AI Usage Policies: Develop and enforce custom internal guidelines and business rules that govern AI model behavior and output, ensuring alignment with organizational ethics and compliance requirements.
- Investigate Specialized AI Security Solutions: Explore platforms like Alice that offer specialized capabilities for AI trust, safety, and security. Such platforms provide dedicated research and intelligence to stay ahead of emerging threats.
- Foster Collaboration: Encourage collaboration between AI development teams and security teams to embed security considerations throughout the AI lifecycle, from design to deployment and ongoing operation. Understanding how to integrate and implement enterprise AI guardrails effectively is critical for minimizing risk in AI-driven environments.
Related: How Agentjacking Exploits AI Coding Agents via Fake Bug Reports, AI Agent Traps: Information as an Attack Surface for Autonomous Systems