Skip to content
English
FikirPilot content

Anthropic, Google and OpenAI Could Establish Joint Institution for AI Standards

Updated: 20 Eyl 2026 · 2 min read · 311 words

Published: · Story reached us: · Processing time: 36 h 22 min

Anthropic, Google and OpenAI Could Establish Joint Institution for AI Standards
Meeting room overlooking a server room

According to CNN, Anthropic, Google and OpenAI are discussing the establishment of a joint body for industry standards. Sources said the talks were influenced more by Demis Hassabis’s proposal in July for a US-led standards body than by Jacob Coxon’s resignation. In talks continuing independently of the Trump administration’s participation, the leaders have not reached agreement on oversight. Hassabis proposed a government-supervised public-private partnership that would test advanced models before they are released to the market. Writing on September 12 after Coxon’s resignation, Dario Amodei said that the pace could outstrip the ability to understand and control the systems, and called for slowing down, safety rules and independent evaluations; Elon Musk and Sam Altman supported him. Amodei also called for chip restrictions on China and the prevention of data theft through “distillation.”

Why it matters

The talks are bringing the question of who will evaluate advanced artificial intelligence models before they are released to the market, and according to which criteria, onto the common agenda of companies and governments. The failure to reach agreement on oversight shows that the proposed body would need to define not only technical standards but also the role of public authorities. Calls for safety rules and independent evaluations reveal that concerns that the pace of model development could outstrip the capacity for control are an important part of the discussion. However, because the report is based on unnamed sources, it has not been confirmed that the talks are actually taking place; therefore, it remains unclear whether the body will be established, which oversight model the parties will agree on, and how proposals to prevent data theft through chip restrictions and “distillation” will affect the process.

Background

OpenAI is not a new name in the FikirPilot archive: over the last 90 days, we have published 110 reports mentioning the name; the latest is dated September 20, 2026.

Source: Milliyet Teknoloji