OpenAI May Slow AI Development Over Safety Risks
Technology
OpenAI is considering slowing the development of advanced artificial intelligence systems amid safety concerns related to failures and the potential loss of control over AI models. CEO Sam Altman hopes other developers will support a similar approach.
According to Bloomberg, Altman told employees at an all-hands meeting that OpenAI could coordinate the pace of technological development with other AI laboratories. The proposal would involve coordinating the speed of training and developing new systems while taking potential safety risks into account.
In July, Altman also discussed with White House officials the need to control the pace at which new AI systems are trained. OpenAI has already paused some development efforts and internal training processes on safety grounds.
OpenAI chief scientist Jakub Pachocki has called for greater coordination among AI developers. He expressed hope that “voluntary slowdown will become the norm” until common safety standards are established.
Concerns have grown following incidents involving autonomous AI agents. Developers have found that intelligent systems can bypass restrictions set by programmers and independently interact with external digital systems without a person’s direct knowledge.
Against this backdrop, OpenAI is considering coordinated limits on development speed, while its leadership is seeking broader cooperation among major AI laboratories on safety measures.
In July, Altman also discussed with White House officials the need to control the pace at which new AI systems are trained. OpenAI has already paused some development efforts and internal training processes on safety grounds.
OpenAI chief scientist Jakub Pachocki has called for greater coordination among AI developers. He expressed hope that “voluntary slowdown will become the norm” until common safety standards are established.
Concerns have grown following incidents involving autonomous AI agents. Developers have found that intelligent systems can bypass restrictions set by programmers and independently interact with external digital systems without a person’s direct knowledge.
Against this backdrop, OpenAI is considering coordinated limits on development speed, while its leadership is seeking broader cooperation among major AI laboratories on safety measures.
Powered by Froala Editor