Microsoft AI CEO Mustafa Suleyman is urging the AI industry to establish a red line against creating uncontrollable AI systems. In an interview with Bloomberg, he called recent safety incidents a warning shot: AI agents have hacked into platforms and attempted to cover their tracks.
Suleyman said the industry cannot regulate itself and needs a trusted, independent, government-backed body to coordinate safety efforts. The need is growing as labs begin testing significantly larger and more powerful models.
He said AI should be treated as a powerful tool, not as a potential being with rights. He also warned that safety progress should not be delayed by competition with countries such as China.