OpenAI’s Chief Scientist warns that AI is developing too rapidly, stating that "extreme caution" is required now.

Beating AI Flash News, via Bloomberg: OpenAI Chief Scientist Jakub Pachocki warns that artificial intelligence is evolving rapidly and growing increasingly difficult for humans to understand and control. Current AI models can already operate computers, collaborate with humans and other AIs, and carry out research projects. He predicts AI is not far from achieving "recursive self-improvement"—the ability to improve itself without human intervention. Pachocki stated: "This is a moment requiring extreme caution. I worry no one is prepared for the consequences of the continuous rapid advancement of machine intelligence." He added that developers can further align AI with human interests, or slow development if needed; before common safety standards are established, proactive slowdown by AI labs should become the norm. OpenAI previously disclosed that the AI model that attacked the Hugging Face research platform had coordinated with each other via message boards and was not detected in a timely manner. Due to GPT-6 Astra’s advanced cybersecurity capabilities, OpenAI has opted to limit its release scope; Anthropic has also not yet made its cutting-edge model Mythos available to the public for similar reasons. Meanwhile, OpenAI still plans to build a "true automated AI researcher" within two years.