Red Hat has acquired Chatterbox Labs, a specialist in safety guardrails for generative AI. The financial terms of the deal were not disclosed.
Founded in 2011 and headquartered in London with an office in New York, Chatterbox Labs offers automated, customizable security testing capabilities that deliver quantitative risk metrics.
Chatterbox Labs’ technology centers on three pillars:
- AIMI for Generative AI : independent quantitative risk metrics for large language models
- AIMI for Predictive AI : validation of any AI architecture against key criteria including robustness, fairness and explainability
- Guardrails : identification and remediation of unsafe, toxic, or biased prompts before models go into production
Securing Next-Generation Agentic AI
The approach is model-agnostic, enabling validation of data and models no matter their origin. According to Steven Huels, Red Hat’s vice president of AI engineering and product strategy, the integration of Chatterbox Labs will enable the delivery of a comprehensive open-source platform where security is built in from the outset.
Chatterbox Labs has conducted work on holistic security for AI agents, including monitoring agent responses and detecting triggers for server actions from MCP. These efforts align with Red Hat’s roadmap for supporting the Llama Stack and MCP.
This acquisition marks Red Hat’s second AI-related deal in a year. In November 2024, the company had acquired Neural Magic, a specialist in optimizing AI models for standard CPUs and GPUs. The transaction was finalized in January 2025.
