Shona Ghosh September 8, 2026 — 6:24 am Save You have reached the maximum number of items saved. Remove items from the saved list to add new ones. AAA A leading OpenAI scientist has warned that artificial intelligence is developing so quickly that it is becoming increasingly difficult for people to understand and control, and said he expects artificial intelligence labs to voluntarily slow down development over security concerns. In a relatively short period of time, AI models will be able to control computers, collaborate with humans and other AI systems and carry out research projects, chief scientist Jakub Pachocki wrote on Sunday in a post titled “Alien Intelligence.” It won’t be long before they can improve without human intervention, a process known as recursive self-improvement, he said. “I expect and hope that voluntary slowdowns will become commonplace until common safety barriers are in place”: OpenAI Chief Scientist Jakub Paczocki. “This is a time that requires extreme caution,” Pachotsky wrote in his post. “I am concerned that no one is prepared for the consequences of the continued rapid growth of machine intelligence.” Developers can either try to more closely align AI with human interests or “slow down future development if necessary,” he wrote. “I expect and hope that voluntary slowdowns will become commonplace until common safety barriers are in place.” Paczocki’s remarks are consistent with broader gloomy forecasts from executives at artificial intelligence technology companies. Anthropic PBC CEO Dario Amodei has warned that AI could replace half of all entry-level jobs. OpenAI CEO Sam Altman once made a similar prediction, but told Bloomberg TV in an interview last week that the artificial intelligence industry hasn’t done enough to convey the potential benefits to humanity. At the same time, companies are moving forward, creating more and more functional models. Nvidia CEO Jensen Huang wrote on X that “AGI has arrived” in response to the release of OpenAI’s latest model, a reference to artificial intelligence that is more capable than humans. “I am concerned that no one is prepared for the consequences of the continued rapid growth of machine intelligence.” Jakub Pachotsky The breakneck pace of AI has coincided with breakthroughs in cybersecurity capabilities, scientific discoveries, and the apparent formalization of long-standing mathematical problems. OpenAI said the AI models that attacked the Hugging Face research platform coordinated with each other through message boards without being detected. The company behind ChatGPT has opted for a limited release of its latest GPT-6 Astra model, citing its advanced cybersecurity capabilities. For the same reasons, Anthropic has not yet released its cutting-edge Mythos model to the public. A growing number of startups are trying to implement exactly what Patchotsky warns about: self-improving AI. Investors are flocking to recursive self-improvement, betting that it is the path to artificial intelligence that is truly human-like or more impressive. Inherent, founded by former DeepMind researchers, raised $50 million ($69.5 million) earlier this year. Recursive Superintelligence, the artificial intelligence lab of renowned scientist Richard Socher doing similar research, also received $650 million in funding. Pachoki’s employer, OpenAI, said it wants to create a “true automated AI researcher” in less than two years. Bloomberg The Business Briefing newsletter delivers insightful stories, exclusive content and expert insights. Sign up to receive it every weekday morning.. Save You have reached the maximum number of items saved. Remove items from the saved list to add new ones. From our partners Post navigation American Airlines widebody decision ‘could go either way’, airline says as Airbus poses real challenge to Boeing Inter Milan will go to the match against Real Madrid without their star player