signal

AI labs must build release processes that satisfy government safety demands

This new regulatory bottleneck forces AI product teams to design release strategies that accommodate uncertain, slow government safety reviews, impacting deployment speed and economic returns while requiring collaboration with independent safety experts.

Frame 1 of 4

It’s not about Anthropic vs. OpenAI anymore

The U.S. government is set to take an awful lot of control over which AI models get released. Two weeks after the U.S. government pulled Anthropic’s Fable and Mythos models, OpenAI’s new model seems to be headed for the same limbo. GPT 5.6 is released only into limited preview, with government approving releases customer by customer. This slows deployment and risks economic upside, forcing AI labs to rethink release processes under regulatory scrutiny. The U.S. government pulled Anthropic’s Fable and Mythos models from general release, limiting them to preview. OpenAI’s GPT 5.6 is being released only into limited preview with government approval 'customer by customer'.