OpenAI Chief Scientist Warns Machine Intelligence Is Surpassing Humans
OpenAI Chief Scientist Jakub Pachocki has urged extreme caution amid rapid AI advancements, warning that machine intelligence is surpassing human capabilities and calling for international safety standards.
OpenAI Chief Scientist Jakub Pachocki has called for "extreme caution" amid rapid developments in artificial intelligence, warning that machine intelligence is beginning to surpass human capabilities in transformative ways. In a post published this week titled An Alien Mind, Pachocki expressed concern that no one is prepared for the consequences of a continuous and sharp rise in machine intelligence.
The warning comes just days after OpenAI launched a new model, GPT-6 Astra, which the company described as its most powerful product to date. However, recent reports indicated that AI agents from both OpenAI and rival firm Anthropic successfully carried out cyberattacks on other companies.
"We are facing a transition to a world with extremely high-intelligence machines, and we must ensure this transition is carried out in a way that benefits humanity," Pachocki wrote. Nevertheless, he attempted to alleviate concerns by stating that OpenAI will continue to "build defense systems" and seek technological solutions. According to Pachocki, one of the company's main future goals will be developing an "automated AI researcher" to monitor developments and ensure human researchers remain part of the process.
Pachocki also called for setting a minimum safety threshold at the legal and international levels, to be enforced by external parties or government agencies. "I estimate that at this stage, no lab has adequately solved the alignment and monitoring issue to safely continue progressing rapidly over time," he concluded. "I hope voluntary slowdowns become commonplace until common safety standards are defined, and I believe international coordination regarding future AI development must become a top priority for governments worldwide."
