AI companies' warnings about recursive self-improvement come as researchers report that increasingly capable AI agents can write software, perform autonomous tasks and assist in developing AI systems. This raises concerns that the technology could advance faster than safeguards to monitor and control it.