Microsoft's AI chief Mustafa Suleyman has issued one of the most aggressive automation warnings yet: lawyers, accountants, project managers, and other white-collar professionals could see most of their tasks fully automated by AI within 12 to 18 months. The timeline far exceeds what most business and policy leaders have anticipated, sending fresh shockwaves through labor markets worldwide.
Microsoft executive sets 18-month replacement timer
In an interview with the Financial Times, Suleyman stated that AI would achieve human-level performance on the “majority, if not all” professional tasks within the next 12–18 months, focusing specifically on desk-based white-collar roles. His prediction is not isolated. Anthropic CEO Dario Amodei issued a similarly stark warning last month, predicting that AI would disrupt 50% of entry-level white-collar jobs within one to five years.
Layoff data is already reflecting the trend. According to Challenger, a job placement firm, 7,624 jobs were cut in January 2026 due to AI, accounting for 7% of total monthly layoffs. For all of 2025, AI-related layoff announcements totaled 54,836. Since tracking began in 2023, 79,449 planned job cuts have been attributed to AI.
White-collar workers train their own replacements: The Mercor model
A particularly ironic case is emerging. The Wall Street Journal reported that Bay Area startup Mercor has “quietly hired tens of thousands of white-collar contractors,” many of them highly qualified professionals in medicine, law, finance, engineering, writing, and art. Their assignment: train the AI systems that may eventually replace them. Contractors earn $45 to $250 per hour reviewing and rewriting model outputs for companies including OpenAI and Anthropic, providing essential data-labeling support for weeks or months at a time.
This model highlights a structural tension in the AI supply chain: the short-term demand for annotation labor clashes with the long-term logic of replacement, destabilizing compensation structures and job security.
Disagreement over the pace of disruption
Not all analysts share the same urgency. Morgan Stanley's global thematic research head Stephen Byrd told clients: “While AI adoption may be faster than past technologies, we think it is still too early to see it in economic data, aside from business investment.” The bank expects the first undeniable impact to arrive in the “late part of this decade and into the next.”
Anthropic CEO lists six AI risks
Dario Amodei's comprehensive warning last month outlined six major risks: mass unemployment (50% of entry-level white-collar jobs gone within 1–5 years); AI with state-level power (envisioning a “genius country” of 50 million people, each more capable than any Nobel laureate, by 2027); rising terrorism threat (biology being the most worrying domain); empowering authoritarians (governments everywhere will have the technology); risks from AI companies themselves (data centers could be used to brainwash consumer bases); and silencing the powerful (AI will generate trillions of dollars annually, making constraint extremely difficult).
Amodei called on “rich people to help solve this problem” and expressed sadness that “many rich people, especially in tech, have recently adopted a cynical and nihilistic attitude, dismissing philanthropy as fraud or useless.”