Microsoft AI Chief Mustafa Suleyman Warns Most White-Collar Jobs Could Be Automated Within 18 Months

Microsoft AI chief Mustafa Suleyman has warned that artificial intelligence could automate most white-collar jobs within the next 12 to 18 months, potentially disrupting professions such as law, accounting, marketing and project management.

In an interview with the Financial Times, Suleyman said Microsoft is working to expand its presence in the enterprise AI market by developing what he described as “professional-grade AGI” — systems capable of performing many tasks traditionally handled by human professionals.

According to Suleyman, these AI tools could automate routine work carried out by employees whose roles primarily involve computer-based tasks.

Sweeping Workforce Changes Ahead

Suleyman predicted significant workforce disruption, arguing that most white-collar roles could be affected in the near term.

“White-collar jobs — those sitting in front of computers, whether lawyers, accountants, project managers or marketers — most of these tasks will be fully automated by AI within the next 12 to 18 months,” he said.

While he did not suggest full job elimination across all sectors, his comments point to substantial automation of task-based work.

AI Model Creation To Become More Accessible

Suleyman also said that building AI systems will become increasingly simple as the technology matures. He compared the future creation of AI models to producing a podcast or writing a blog, suggesting that both organisations and individuals will be able to design models tailored to their needs.

Looking ahead two to three years, he added that AI agents could become significantly more efficient at managing workflows across large enterprises and institutions.

Push Toward AI Self-Sufficiency

The Microsoft AI chief indicated that the company intends to strengthen its in-house AI capabilities and reduce reliance on OpenAI following adjustments to their partnership structure.

He said Microsoft is aiming for what he called “true AI self-sufficiency,” with internally developed AI models potentially debuting as early as 2026.