Three leading artificial intelligence companies are exploring the creation of an industry standards organization that would establish shared criteria for assessing advanced AI models before public release, according to reporting by CNN. Anthropic, Google, and OpenAI have been in talks about forming this body, though discussions remain ongoing and no consensus has emerged on what specific requirements or safeguards the standards would include. The initiative could eventually provide technology partners and enterprise buyers with a more uniform method for evaluating AI models ahead of deployment.
The conversations were initiated by a July proposal from Demis Hassabis, co-founder and CEO of Google DeepMind, who suggested creating a US-led standards body modeled after the Financial Industry Regulatory Authority, known as FINRA. CNN reported that Hassabis envisioned an organization that would examine advanced AI models before wide-scale deployment, potentially operating as a public-private partnership with government oversight, industry funding, and staffing from independent technical specialists and open-source community representatives. OpenAI has separately engaged with external organizations to explore what concrete standards might entail, and CEO Sam Altman has voiced public support for collaborating with rival AI companies on the effort. "I believe that shared safety standards and international coordination on further AI development need to be priorities now," OpenAI chief scientist Jakub Pachocki told reporters earlier this month, according to CNN, adding that OpenAI expects to share more details about potential standards in coming months.
These industry talks run parallel to a separate proposal under consideration by the Trump administration. Treasury Secretary Scott Bessent helped craft a plan for an independent AI regulator that would assess the safety of advanced models with input from industry, Bloomberg reported in July. That proposed entity would report to the Securities and Exchange Commission and follow a structure resembling FINRA. A June executive order directed federal agencies to design a voluntary framework allowing developers to grant government access to designated frontier models for up to 30 days before releasing them to other trusted partners, though the order doesn't mandate government approval before release and hasn't publicly detailed the benchmarking or designation process. Not all major tech companies back the idea of a national AI regulator—Meta CEO Mark Zuckerberg reportedly opposed the concept during a summer call with President Donald Trump, describing it as fundamentally flawed.
A shared testing framework could offer resellers, integrators, consultants, and other channel partners an additional method for comparing AI products from competing vendors, according to the report. Partners currently must weigh differences in model performance, security measures, deployment needs, data management, and governance when recommending AI platforms to clients. If major vendors begin publishing results under the same framework, agreed-upon testing standards could make those evaluations more consistent. Enterprise customers might start asking whether AI products satisfy recognized standards in the same way they currently assess security certifications, compliance obligations, and vendor risk. For managed service providers, the development could prove especially significant as more customers seek help choosing, deploying, and managing AI tools—common standards could give MSPs another metric for deciding which models to support and for clarifying the risks and distinctions between competing platforms. Whether this initiative produces comparable test results, independent oversight, or recognized certifications that customers incorporate into procurement requirements depends entirely on whether Anthropic, Google, OpenAI, and other AI companies can reach agreement on common standards. Channel partners don't need to alter their model-selection processes immediately, but monitoring these discussions makes sense as the industry navigates how to balance innovation with accountability. The path forward hinges on whether competitive pressures give way to collaborative frameworks that serve both vendor interests and customer expectations.