In a move to address growing security concerns around autonomous AI, OpenAI confirmed on Monday the acquisition of Promptfoo, a startup specializing in testing large language models for vulnerabilities. The deal will see Promptfoo's technology folded into OpenAI Frontier, the company's platform for business-oriented AI agents.

As AI agents that execute complex digital tasks become more common, their potential for productivity is matched by new risks. Malicious actors can exploit weaknesses to steal data or hijack automated processes. This acquisition signals a major effort by OpenAI to harden its systems for high-stakes corporate environments.

Founded in 2024 by Ian Webster and Michael D’Angelo, Promptfoo built an open-source framework and tools for stress-testing LLMs against security threats. The startup claims its software is already deployed in over a quarter of Fortune 500 companies. Despite its market penetration, Promptfoo remained lean, having raised only $23 million. Its valuation stood at $86 million as of last July.

OpenAI stated that integrating Promptfoo will enable automated security testing, risk assessment for agent workflows, and continuous monitoring for compliance. The company also plans to maintain and expand Promptfoo's open-source projects. Financial terms of the transaction were not disclosed.

Source: TechCrunch