AI researchers warn companies rushing self-improving systems despite safety risks

Current and former OpenAI and Google DeepMind researchers said labs are racing to develop self-improving AI despite concerns the technology could outpace human control.