Story perspectives

OpenAI Won’t Release GPT-6.1 Astra After Safety Failure

By Drooid · Sep 29, 2026, 6:08 PM UTC · How we work

Summary

Story summary

U.S. House Speaker Mike Johnson urged voluntary AI guardrails in a Tuesday meeting with President Donald Trump and CEOs.
Johnson said products liability law will apply to AI and liability exemptions “don’t come up”.
OpenAI announced it will not release GPT-6.1 Astra because the model failed its safety standard.
Anthropic warned its technology poses “catastrophic or existential risk to humanity” in its IPO prospectus.

History

Historical Context

The EU adopted the General Data Protection Regulation on 14 April 2016, with its rules taking effect on 25 May 2018 to give individuals control over personal data. Enforcement quickly followed, exemplified by the French data-protection authority’s €50 million fine against Google on 21 January 2019. U.S. officials such as Speaker Mike Johnson now cite those penalties while urging similar AI liability rules.

1 of 2

Story summary

U.S. House Speaker Mike Johnson urged voluntary AI guardrails in a Tuesday meeting with President Donald Trump and CEOs.
Johnson said products liability law will apply to AI and liability exemptions “don’t come up”.
OpenAI announced it will not release GPT-6.1 Astra because the model failed its safety standard.
Anthropic warned its technology poses “catastrophic or existential risk to humanity” in its IPO prospectus.

1 / 2