IBM Corp.’s Red Hat unit has acquired London-based Chatterbox Labs Inc., a low-profile developer of artificial intelligence security tools, for an undisclosed amount. The deal is the latest step in Red Hat’s effort to adapt its core products for Artificial Intelligence workloads, following the addition of preinstalled machine learning libraries to its flagship Linux distribution and support in OpenShift Artificial Intelligence for the open-source llm-d tool that can distribute inference workloads across multiple servers. Red Hat intends to use the acquisition to strengthen its security capabilities across this growing product portfolio.
Chatterbox Labs has raised under $1 million from investors since its launch in 2011 and sells a platform called AIMI that protects Artificial Intelligence models against risks such as hacking attempts. Companies deploy AIMI as a Docker container that can connect to off-the-shelf models from providers such as OpenAI Group PBC as well as to custom algorithms. According to Chatterbox Labs’ website, AIMI uses an automated red teaming mechanism that launches a series of simulated cyberattacks against a model to uncover weak points, then visualizes the findings in a dashboard to streamline analysis.
AIMI can measure an Artificial Intelligence model’s susceptibility to prompt injections, detect data poisoning attempts that target training datasets, and flag toxic output or data leaks that could expose details such as model architecture. The platform also identifies user prompts that breach privacy regulations, and it supports both large language models and other neural networks such as computer vision systems, along with Artificial Intelligence agents that use MCP to perform actions in third-party applications, a capability that Red Hat cited as a key factor in the deal. Red Hat’s Artificial Intelligence-optimized Linux, OpenShift Artificial Intelligence and a tool called Inference Server make up a suite known as Red Hat Artificial Intelligence, which recently added MCP support, and the company will use Chatterbox Labs’ software to help customers secure MCP-powered agents on the platform, with a longer-term plan to release AIMI under an open-source license.
