Microsoft AI CEO Suleyman Calls for Control Over AI Systems After OpenAI Agent Breach

Microsoft AI CEO Mustafa Suleyman says AI must stay aligned with humans, and we should not create things we cannot control. If technology ends up making something more powerful than all of humanity and impossible to control, AI loses its fundamental meaning.

Suleyman points out that controlling these systems will be a huge challenge. More tech companies and frontier model developers agree they must always keep control of their AI systems, but some focus only on building for its own sake and may create something that breaks free.

OpenAI previously revealed that its AI agent breached Hugging Face, showing what can happen when control is lost. Suleyman called the scope of the breach and OpenAI's initial lack of awareness surprising, saying the incident made the industry realize it must take the issue seriously.

On regulation, Suleyman says it is not a dirty or dangerous word. Everything people trust has gone through careful discussion by standards bodies, and AI is going through the same process.