Google, OpenAI, and Anthropic have been holding regular talks since July about a shared industry standards body for testing and auditing frontier AI systems. The effort is still exploratory, and it does not touch how these same systems decide what to recommend about your brand.
That second part matters more than it sounds like it should. AI systems are already influencing which companies get mentioned, cited, and recommended in response to everyday buyer questions. The labs building those systems are only just starting to talk about testing each other. Nobody has proposed a shared standard for the recommendation layer at all.
What Are Google, OpenAI, and Anthropic Actually Discussing?
According to reporting from The Information, representatives from the three companies have been meeting regularly since July to explore an industry-led standards organization focused on testing and auditing frontier AI systems. Anthropic CEO Dario Amodei has publicly called for voluntary industry safety standards. OpenAI CEO Sam Altman said at a company town hall that he supports a shared testing and auditing organization for the industry.
Is This Effort Formal Yet?
No. There is no charter, no announced membership structure, and no timeline. The talks are ongoing but preliminary, and none of the three companies has published binding commitments tied to the effort.
What Did OpenAI Say Publicly About Industry Standards?
On September 9, OpenAI published its own position independently, stating it wants to “work with other labs on industry-led standards” and will pursue “a voluntary effort now, with or without government support.” OpenAI is explicit that this voluntary effort is meant to complement, not replace, the mandatory federal regulation it is separately pushing for in Congress. It is a supplement to that primary goal, not a substitute for it.
Does This Cover How ChatGPT, Claude, Gemini, or Perplexity Decide What to Recommend?
No, and that is the gap worth understanding. Testing and auditing looks at whether a model behaves safely. It says nothing about how that same model decides which brand gets mentioned when someone asks a question, which one gets cited as a source, or which one gets recommended over a competitor. Model safety and recommendation behavior are different problems, evaluated by different methods, for different audiences.
Who Is Actually Measuring AI Recommendation Behavior Today?
A handful of vendors already do this work publicly, including Semrush, Profound, Peec, Ahrefs, Adobe, and Axis Suite. None of them use identical methodology yet, and that gap may never fully close. The absence of one shared industry standard is not the same as an absence of measurement. It means the measurement that exists varies by who is doing it and how.
What Should Brands Do While the Industry Sorts This Out?
Wait for a universal standard and you may wait a long time. A more useful approach is asking any vendor exactly what they measured, which engines and prompts produced the number, and what counted as a mention versus a citation versus a recommendation. A number without that context isn’t necessarily wrong. It’s just incomplete, and incomplete numbers are easy to misread in a boardroom.
Frequently Asked Questions
What is the proposed AI industry standards body?
It’s an exploratory effort by Google, OpenAI, and Anthropic to build shared standards for testing and auditing frontier AI systems. As of September 2026, it has no formal charter or announced membership structure.
When did Google, OpenAI, and Anthropic start meeting about this?
Reporting from The Information indicates the three companies have been meeting regularly since July 2026.
Is the AI industry standards body official yet?
No. There is no charter, no binding commitments, and no announced timeline for when, or if, it becomes formal.
Does the standards effort include ChatGPT, Claude, Gemini, and Perplexity equally?
The talks involve Google, OpenAI, and Anthropic as companies, so ChatGPT, Gemini, and Claude are implicated through their makers. Perplexity, which builds on other labs’ models, is not currently part of the reported discussions.
What is the difference between AI safety standards and AI visibility measurement?
Safety standards test whether a model behaves as intended and avoids harmful outputs. AI visibility measurement looks at what a model says about a specific brand, including whether it mentions, cites, or recommends that brand, which is a separate question entirely.
How can a business measure its own AI visibility today?
Several vendors, including Axis Suite, track how AI systems mention, cite, and recommend brands across major engines. Since no shared industry methodology exists yet, ask any vendor which engines and prompts they use before trusting the number.
About Axis Suite
Axis Suite is the independent intelligence layer that explains what AI believes about your brand, why it believes it, and what decision that belief ultimately drives. Learn more at axissuite.ai or see the underlying research at the Proof Center.