A report by the Future of Life Institute reveals that major AI companies, including OpenAI and Google DeepMind, face significant safety vulnerabilities in their models. While Anthropic ranked highest with a C grade, other companies like Meta and x.AI received low grades (F and D- respectively) for inadequate safety measures. The report emphasizes the need for independent oversight and suggests that current safety strategies are insufficient to ensure future AI systems remain under human control.