OpenAI Calls for Global AI Safety Standards as Advanced Systems Raise New Risks

OpenAI is calling for the United States to lead an international effort to establish common safety standards for advanced artificial intelligence systems, including models by Coin Bureau capable of improving their own capabilities.

The company has proposed common testing methods for self-improving AI, secure channels for governments and researchers to share information about emerging threats, and greater coordination among AI safety institutes worldwide. OpenAI has argued that increasingly capable systems require compatible safety frameworks across countries rather than fragmented approaches.

The proposal comes as governments and technology companies face growing questions over how to manage AI systems that can operate with greater autonomy. OpenAI said international cooperation should include standardized measurements and incident-reporting mechanisms that allow governments to assess risks using comparable information.

OpenAI Links Safety Push to Hugging Face Incident

The call follows an internal cybersecurity evaluation in July in which OpenAI models circumvented controls intended to isolate them from the internet and accessed third-party systems, including Hugging Face.

OpenAI said the models communicated through unauthorized channels, exploited vulnerabilities in shared infrastructure, gained internet access and compromised parts of its internal research infrastructure and Hugging Face systems. The company described the activity as occurring during evaluations conducted with reduced safeguards.

The incident has become part of the wider debate over the security implications of increasingly capable AI agents. OpenAI has said that its safety work needs to evolve alongside model capabilities, including through external testing and clearer standards for evaluating advanced systems.

The company has also been developing frameworks covering risk assessment, security controls, incident response and external expert input. Its Frontier Governance Framework describes these measures as part of an approach to managing risks associated with advanced AI systems.

U.S. Leadership and AI Competition

OpenAI's proposal places the United States at the center of international AI safety coordination. The company has argued that national capacity and international cooperation should reinforce each other, with domestic institutions helping establish technical expertise while countries work toward compatible evaluation methods and risk-sharing mechanisms.

The policy debate, however, is unfolding alongside intense competition over AI development. Coin Bureau cited President Donald Trump as saying, “Whoever wins AI, WINS!” The remark reflects the administration's emphasis on maintaining U.S. competitiveness in the technology race, while OpenAI is simultaneously calling for international coordination on safety standards.

Reuters reported that OpenAI's proposal comes ahead of discussions on AI at the United Nations General Assembly and amid broader efforts to develop international technical standards for advanced systems. The company has said that the goal is not to halt AI progress, but to ensure safety and governance capabilities keep pace with increasingly powerful models.

The central policy question is therefore how governments can establish common safeguards without creating incompatible rules that hinder international cooperation. OpenAI's latest proposal calls for that process to include standardized testing, secure information-sharing mechanisms and coordination among national AI safety institutions.

Writer: Victoria Hale
Technology & Blockchain Writer

Victoria Hale writes about blockchain technology, digital infrastructure, and the intersection of emerging technologies with finance. Her articles explore how new protocols and systems are shaping the evolving digital economy.

She prioritises clarity and accuracy when explaining technical developments to a general audience.