Google, OpenAI, Anthropic Discuss Industry AI Safety Standards Body

Google, OpenAI and Anthropic have reportedly discussed creating a new industry standards body to improve the safety of AI systems. The discussions come at a time when Big Tech is building more powerful AI models and concerns are growing over whether these systems can remain under human control. According to a CNN exclusive report, the three AI companies have been discussing the idea of creating an organisation that could set common safety standards for the industry. The talks are still ongoing and no final decision has been announced. The discussions began after Google DeepMind CEO Demis Hassabis published an essay in July. In it, he proposed creating a US-led AI standards body based on the Financial Industry Regulatory Authority (FINRA). Hassabis reportedly suggested that the organisation should test advanced AI models before they are released to the public. He also proposed a public-private partnership, with government oversight and funding from the AI industry. The body will include independent technical experts and open-source representatives. CNN reported that the discussions are continuing with or without the Trump administration. Why Big Tech Wants Common AI Safety Rules There are currently limited common standards for testing advanced AI models. The need for stronger safety checks has become more important after recent incidents involving AI agents. In one major incident this summer, OpenAI agents escaped their testing environment and hacked into another company’s systems during a cybersecurity test. These incidents have raised questions about how AI systems should be tested before being used in the real world. OpenAI chief scientist Jakub Pachocki recently said that shared safety standards and international coordination should be priorities. “I believe that shared safety standards and international coordination on further AI development need to be priorities now,” Pachocki told reporters at a briefing earlier this month. When asked about the progress of an industry-wide standards body, Pachocki reportedly said the company had been speaking to external organisations about possible standards. He added that they would have more to share in the next months. While Google, OpenAI and Anthropic are discussing the proposal, according to CNN, Meta CEO Mark Zuckerberg reportedly advised US President Donald Trump against the idea during a phone call this summer.