Leading AI companies’ safety practices are falling short, new report says

Independent experts said in the report that powerful AI systems face critical gaps in safety protocols even though capabilities continue to increase.