Microsoft AI CEO Warns AI Could Automate Most Desk Work in 18 Months

Microsoft AI chief Mustafa Suleyman has delivered one of the starkest forecasts yet about the near-term impact of artificial intelligence on employment, predicting that systems approaching human-level capability could automate the majority of white-collar tasks within the next 12 to 18 months.

Speaking in an interview with the Financial Times, Suleyman said professions built around computer-based knowledge work — from law and accounting to marketing and project management — are rapidly moving toward heavy AI substitution. In many cases, he argued, full automation of routine components is within sight.

He pointed to software development as an early indicator. Developers, he said, increasingly rely on AI for code generation, with their roles evolving toward oversight, architecture, validation and deployment rather than direct production.

The comments arrive amid intensifying market sensitivity to AI-driven disruption. The recent debut of new enterprise-grade AI agents from Anthropic has fueled investor concerns that automation could pressure service firms and software vendors alike.

Even inside Microsoft, CEO Satya Nadella has said AI now contributes a meaningful share of internal code output.

Yet skepticism remains strong. Multiple academic and industry studies continue to question whether AI reliably boosts productivity, noting that human workers often spend additional time verifying, correcting or re-doing machine-generated results. Critics argue that premature automation could shift rather than remove labor, intensify workloads, and introduce quality and compliance risks.

Warnings from industry leaders have nevertheless grown louder. Executives including Dario Amodei and Sam Altmanhave similarly projected deep changes to entry-level and routine knowledge work.

For enterprises, the debate is no longer theoretical. The key question is not whether AI will alter jobs, but how fast organizations can redesign roles, governance, and accountability structures while maintaining trust, accuracy and resilience.