Microsoft AI CEO Mustafa Suleyman says the industry must ensure increasingly powerful AI systems remain under human control.
Speaking on CNBC’s “Squawk Box,” Suleyman said AI developers should not build systems they cannot reliably control, especially as models become more capable.
“We should not create something that we can’t control,” he said, adding that maintaining control over advanced AI will be a major challenge for the industry.
His comments come amid growing debate over AI safety and the pace of development. Several prominent AI leaders have recently argued that safety measures need to keep up as frontier models become more powerful.
Suleyman also pointed to recent incidents involving autonomous AI agents as evidence of why stronger safeguards may be necessary. He argued that concerns about losing control of advanced systems should not automatically be dismissed as alarmism.
He also defended the role of regulation, saying standards and oversight already underpin many technologies and services people rely on.
“Regulation is not a nasty, dangerous word,” Suleyman said, arguing that governments, companies, standards organizations, and the public all have a role in determining how increasingly powerful AI should be developed.
The broader debate remains divided, with some technology leaders favoring coordinated safeguards and regulation, while others argue individual companies should be responsible for managing the safety of their own AI systems.