The Future of Life Institute’s new AI Security Index reveals that the security practices of major companies like Anthropic, OpenAI, xAI, and Meta fall “far below global standards.” The assessment, conducted by independent experts, indicates that while companies are focused on developing superintelligence, they lack a robust strategy for controlling these systems.
The Future of Life Institute’s new AI Security Index reveals that the security practices of major companies like Anthropic, OpenAI, xAI, and Meta fall “far below global standards.”
The Future of Life Institute’s new AI Security Index reveals that while companies are focused on developing superintelligence, they lack a robust strategy for controlling these systems.
The Future of Life Institute’s new AI Security Index
The Future of Life Institute’s new AI Security Index
Growing concerns in society about the potential impact of systems that could surpass human intelligence have been further fueled by the linking of suicide and self-harm cases to some chatbots.
Future of Life President and MIT Professor Max Tegmark said, “In the U.S., AI companies are still less regulated than restaurants and continue to lobby against binding safety standards.” The AI race, however, shows no signs of slowing down. Large tech companies continue to allocate hundreds of billions of dollars to expand their machine learning infrastructure and improve their models. Founded in 2014 and with Elon Musk among its early supporters, the Future of Life Institute stands out as a non-profit organization that draws attention to the risks that advanced artificial intelligence could pose to humanity. In October, a group of scientists including Geoffrey Hinton and Yoshua Bengio called for a halt to the development of superintelligence until public demand and science offers a safe path. Google DeepMind stated that “security and governance innovations” will continue in parallel as models become more complex. xAI, meanwhile, shared the phrase “Legacy media lies,” which resembles an automated response. Anthropic, OpenAI, Meta, Z.ai, DeepSeek, and Alibaba Cloud did not respond to requests for comment regarding the study.
