Microsoft AI chief says white-collar work could be fully automated within 18 months
By Staff Writer
Microsoft’s top artificial intelligence executive has issued one of the starkest predictions yet about the speed of workplace disruption, saying that most professional white-collar tasks could be automated by AI within the next 12 to 18 months.
Mustafa Suleyman, the CEO of Microsoft AI, made the comments in a recent interview with the Financial Times, arguing that AI systems are rapidly approaching human-level performance across a wide range of office-based work. His remarks add fresh urgency to the debate over how quickly artificial intelligence will reshape jobs in law, accounting, marketing, project management and other knowledge-based fields.
“I think that we’re going to have a human-level performance on most, if not all, professional tasks,” Suleyman said, according to reporting cited by multiple outlets. “So white collar work, where you’re sitting down at a computer, either being a lawyer or an accountant or a project manager or a marketing person, most of those tasks will be fully automated by an AI within the next 12 to 18 months.”
A bold timeline from inside Big Tech
The prediction stands out not only for its scope, but for its timeline. Many AI researchers and labor economists have warned that generative AI will eventually transform office work, but Suleyman’s estimate suggests a much faster transition than many businesses or workers may be prepared for.
Microsoft has been one of the most aggressive corporate adopters of AI in the enterprise market, integrating its technology into software products used daily by millions of workers. Suleyman’s remarks reflect a broader push across the technology industry to position AI as a tool that can handle increasingly complex tasks, rather than simply assist human employees.
His comments also underscore a growing shift in how AI is being framed. Earlier waves of automation were often described as replacing repetitive, manual labor. The latest generation of AI, however, is now being presented as capable of handling the analysis, writing, organization and communication that define many office jobs.
What jobs could be affected?
According to Suleyman’s outlook, the most exposed roles are those built around computer-based work. That includes professions where much of the day is spent drafting documents, reviewing data, preparing presentations, coordinating workflows or communicating with clients.
Fields such as accounting, legal services, marketing, consulting and project management could be among the first to feel the effects if AI systems continue improving at their current pace. In those industries, even partial automation could reshape hiring, reduce routine workloads and force workers to shift toward oversight, strategy and client-facing responsibilities.
For many companies, the appeal is obvious: AI can potentially operate around the clock, cut costs and speed up production. For workers, however, the threat is less abstract. If routine white-collar tasks can be done faster and cheaper by software, the pressure to redefine human value in the workplace will intensify.
Growing debate over AI and employment
Suleyman’s remarks arrive at a moment when CEOs, policymakers and workers are already grappling with AI’s impact on employment. Tech leaders frequently stress that AI will create new roles even as it eliminates others, but critics argue that the transition could be turbulent and uneven, especially for entry-level and mid-tier knowledge jobs.
Some economists say that previous waves of technology ultimately created more jobs than they destroyed, though not always for the same workers or in the same locations. The concern today is that AI may move too quickly for retraining systems, labor markets and education pipelines to adapt.
That fear is particularly acute in white-collar sectors, where professionals once assumed their work was safer from automation than factory or warehouse jobs. AI is now challenging that assumption by targeting tasks that involve text generation, research, analysis and decision-making support.
Beyond efficiency: questions about trust and control
Suleyman has also warned about the safety implications of rapidly advancing AI. In related discussions, he has suggested that the world may soon face serious AI-related incidents and that current governance systems may not be ready to manage them. Those concerns add another layer to the automation debate, shifting it from jobs alone to the broader question of how much autonomy these systems should be given.
As AI becomes more capable, businesses will face difficult choices about whether to deploy it broadly, restrict it to assistive functions or use it to replace human labor in specific departments. The more AI is trusted to act independently, the more important it becomes to ensure accountability, transparency and human oversight.
For workers, the immediate challenge may be less about a sudden disappearance of jobs and more about the steady erosion of tasks that once defined their role. If AI handles first drafts, data summaries, scheduling, research and client correspondence, human employees may find themselves managing exceptions rather than performing the core work themselves.
What comes next
Whether Suleyman’s 18-month forecast proves accurate remains to be seen. AI capabilities are improving rapidly, but real-world adoption often depends on reliability, regulation, cost and organizational culture. Many businesses will be cautious about handing mission-critical tasks to systems that can still make mistakes.
Even so, the warning from one of the most prominent executives in AI development is likely to intensify the conversation inside boardrooms and government offices. Companies may accelerate investment in AI training and workflow redesign, while workers may seek reassurance that their roles will evolve rather than vanish.
At minimum, Suleyman’s prediction reinforces a message that has become harder to ignore: the future of office work may arrive faster than expected, and the people most affected may not be those who assumed automation was someone else’s problem.