OpenAI and Anthropic: AI Alert Amid Rapid Innovations

OpenAI and Anthropic, two of the leading artificial intelligence laboratories in the world, have recently released a series of policy papers warning about the risks associated with the rapid development of AI. These publications come as both companies continue to launch new AI models at a brisk pace, including GPT-5.5 and Claude Fable 5.

Over the past two months, these labs have ramped up their research and policy articles, highlighting an AI development that seems out of control, while government regulations struggle to keep up. This situation led Anthropic to publish a document last week, calling for a coordinated "slowdown or pause" on a global scale in the development of cutting-edge models. According to Anthropic, without a global coordination mechanism, companies and governments will face difficult decisions regarding safety, under competitive and geopolitical pressure.

Anthropic's CEO, Dario Amodei, reiterated this idea in a blog post, stating that AI is evolving at a "breakneck pace" while policies progress "very slowly." For his part, OpenAI CEO Sam Altman, along with Chief Scientist Jakub Pachocki, proposed the creation of an "international organization" to coordinate AI efforts and reduce catastrophic risks. This organization should, according to them, slow down the development of cutting-edge AI so that "societal resilience, safety, and alignment can keep pace."

These concerns were expressed as both companies introduced new AI models. On Tuesday, Anthropic launched Claude Fable 5, its most advanced model in the "Mythos" class, which includes safeguards to redirect sensitive requests to the Opus 4.8 model or other solutions. OpenAI also introduced GPT-5.5 at the end of April, touted as the "smartest and most intuitive model" to date, capable of understanding and assisting with user work.

In parallel, both companies have encouraged rapid adoption of their technologies by offering free usage benefits for Codex and Claude Code, tools that allow businesses to create AI agents to automate workflows and enhance productivity.

These warnings come as OpenAI and Anthropic prepare for their highly anticipated IPOs. OpenAI announced on Monday that it had confidentially filed its S-1, about a week after Anthropic. Representatives from both companies did not respond to requests for comment from Business Insider.