Microsoft AI head Mustafa Suleyman has called for strict industry red lines and government oversight, arguing that developers must immediately shut down models that autonomously develop recursive self-improvement or independent goal-setting capabilities.