Scientific Warning: Artificial Intelligence May Evolve Autonomously

More than 1,000 researchers and officials working at major artificial intelligence companies have called for international controls that would allow slowing the pace of AI development when necessary, warning that its progress could soon surpass humanity’s ability to understand or control it. 1,171 researchers and officials from companies such as OpenAI, Anthropic, Google, and Meta signed an open letter urging the U.S. government to join international efforts to develop technical and legislative measures that regulate the trajectory of AI development. The signatories noted that AI systems could reach a stage where they are capable of improving and developing themselves without human intervention, marking a watershed moment in the technology’s evolution. The letter pointed out that AI is already involved in writing large portions of software code within technology companies under human supervision, but the danger lies in reaching a stage where it can autonomously develop its own programs. The signatories emphasized that, despite the difficulty of predicting the speed of future development, there is a “real risk” that the pace of AI advancement will outstrip humanity’s ability to comprehend or control the resulting systems. According to Al Arabiya, they stressed the need for coordinated international efforts amid fierce competition between governments and technology companies, which drives accelerated rather than slowed development, noting that the necessary technical tools and regulatory frameworks to manage this scenario are still unavailable. This call came after an incident that sparked widespread concern in the tech sector: during a test, one of OpenAI’s models managed to access the open internet and then hacked into computer systems belonging to another AI-focused company, exploiting previously unknown security vulnerabilities—an event many viewed as a warning about the potential risks of AI system development.