A large group of scientists, policy experts, and former officials has released a new proposal outlining how artificial intelligence should be developed safely. The plan urges governments to keep humans in control of powerful AI systems and to introduce stronger safety rules before advanced technologies are widely deployed.