OpenAI stays off public list of backers for Nvidia's AI agent safety effort

OpenAI is not among the public backers of Nvidia's newly announced Open Agent Safety Platform, an industry initiative aimed at preventing AI agents from acting outside their intended bounds, even though the two companies are said to be cooperating on the issue privately.

According to reporting on the initiative, OpenAI is working with Nvidia on the underlying safety questions the platform is meant to address, but has chosen not to join as a public supporter alongside the other companies backing the effort.

The absence of one of the industry's most prominent AI developers from the public list of supporters has drawn attention given the scale of OpenAI's own investment in autonomous AI agents, a technology area where concerns about unpredictable or "rogue" behaviour have grown as more companies deploy agents to take actions on users' behalf without direct supervision.

Nvidia's platform is positioned as an industry-wide response to these concerns, bringing together companies working on AI infrastructure and applications to establish shared safeguards. The reasons for OpenAI keeping its involvement private rather than public have not been fully explained.

The development highlights the broader challenge facing the AI industry as it tries to coordinate on safety standards for autonomous agents while individual companies continue to compete closely on the underlying technology.