SydeLabs, a California-based startup, has secured $2.5 million in seed funding to develop an intent-based firewall guard specifically designed for AI systems. The company aims to address the potential risks and vulnerabilities that come with generative AI technology by providing a comprehensive suite of solutions for developers.
The funding round was led by RTP Global, Picus Capital, and other angel investors. With this capital injection, SydeLabs plans to focus on research and development, as well as upgrading its technology stack to stay ahead of malicious actors seeking to compromise generative AI systems.
SydeLabs offers three key AI-focused products: SydeBox, SydeGuard, and SydeComply. SydeBox is currently available in beta and serves as a self-service red-teaming solution that allows teams to stress-test their AI applications and models for vulnerabilities. SydeGuard, which is slated to go live soon, provides real-time intent-based protection and identifies gaps in AI systems that could lead to compliance issues. SydeComply is a compliance-centered offering that aims to flag any violations of internal or external regulations.
Ankita Kumari, co-founder of SydeBox, explained that the red teaming solution uses an AI agent to craft test attacks while an LLM (large language model) detects the success of these attacks based on the response of the target system. On the other hand, SydeGuard uses proprietary models to identify the intent of end-users and assess potential threats. It then shares this data with enterprise security teams, giving them control over how they want to handle a potential attack.
Since its launch on March 1, 2024, SydeBox has been adopted by over 15 enterprises, resulting in the detection of more than 10,000 vulnerabilities across 50+ applications and models. The startup intends to make SydeGuard and SydeComply available to customers in the near future and plans to monetize SydeGuard using a consumption-based model.
The funding for SydeLabs comes at a critical time when the safety and reliability of generative AI applications are under scrutiny due to prompt injection attacks and the creation of deepfakes. Other companies, such as Microsoft and Skyflow, have also announced initiatives to address these concerns.
In the red teaming and real-time prevention space, SydeLabs faces competition from players like Lakera and Prompt Security. However, Kumari claims that SydeLabs’ products outperform these tools in terms of accuracy and performance.
Overall, SydeLabs’ intent-based firewall guard is a promising solution for businesses looking to secure their AI systems. With its suite of AI-focused products and its commitment to continuous improvement and staying ahead of malicious actors, SydeLabs is positioning itself as a leader in the AI security space.