Anthropic and OpenAI are increasingly warning that AI development is advancing faster than governments can regulate it — even as both companies continue releasing more powerful models at a rapid pace.
The two leading AI labs have recently published a series of research papers and policy proposals highlighting concerns about frontier AI progressing faster than safety frameworks and public institutions can adapt.
Last week, Anthropic called for an internationally coordinated slowdown in advanced AI development, arguing that governments need time to establish effective oversight before capabilities accelerate further.
Anthropic CEO Dario Amodei echoed that message, saying AI is advancing at a "lightning pace" while policy responses remain slow.
OpenAI leaders have expressed similar concerns. In a recent blog post, Sam Altman and Jakub Pachocki argued for an international body that could coordinate AI development and help reduce catastrophic risks as systems become more capable.
At the same time, both companies are accelerating deployment. Anthropic recently launched Claude Fable 5, while OpenAI continues expanding access to its latest models and AI-agent tools designed to automate increasingly complex tasks.
The contrast is striking: the companies leading the AI race are also among the loudest voices warning about the risks of moving too fast.