AIceberg announced on Thursday the launch of its AI trust platform, for which it raised $10 million in seed funding from SYN Ventures and Sprout & Oak.
Founded in 2022, the New York-based company has developed a platform designed to help governments and enterprises with the safe, secure and compliant adoption of AI, specifically generative AI and agentic AI.
AIceberg says its platform acts as a firewall and gateway that monitors AI usage for potential risks and enables the enforcement of security and other policies.
The platform is designed to prevent unsanctioned, unsuitable or illegal content, and automatically redacts personal and other sensitive information. It can detect common cyberattack vectors such as prompt injection, prompt leaking, and jailbreaking, and is capable of performing a security analysis of agentic workflows.
The AIceberg platform also provides compliance, transparency, auditability, and observability capabilities.
“Organizations adopting generative and agentic AI face critical challenges in ensuring safety, security, and compliance,” said Alex Schlager, CEO at AIceberg. “Additionally, many are operating with a false sense of security because while AI TRiSM solutions powered by LLMs seem convenient, using LLMs to safeguard LLMs introduces systemic risks and architectural limitations that undermine their effectiveness.”
“We are closing the security gap with AIceberg. Purpose-built, non-generative models to detect risk signals and support safe AI adoption, AIceberg works independently of AI applications, using the content of input and output to detect and eliminate risks and power safe, secure, compliant use of generative models and agentic workflows across the enterprise,” Schlager added.
Related: Blockaid Raises $50 Million to Secure Blockchain Applications
Related: Singulr Launches With $10M in Funding for AI Security and Governance Platform
Related: Ex-NSO Group CEO’s Security Firm Dream Raises $100M at $1.1B Valuation
Related: Astra, Invary Raise Millions for AI-Powered Pentesting, Runtime Security
About The Author
Original post here
