What happened
MIT's 'Iceberg Index,' published November 2025, used AI agent-based simulations on the ONET database, mapping 32,000 skills across 900+ US occupations. Incorporating tools like Microsoft Copilot and Anthropic Claude, it found AI could replace 11.7% of the US labour market, particularly in technology, cognitive, and administrative roles; coding jobs showed high overlap. Separately, Microsoft (February 2026) and Anthropic (March 2026) released frameworks, also using ONET, identifying 'information work' and roles like computer programmers and customer service representatives as highly exposed.
Why it matters
This quantified data provides procurement teams and HR strategists with granular insights into AI's immediate impact on specific job functions and tasks. The mechanism involves three distinct models offering data on task-level AI applicability and job exposure, grounded in the comprehensive ONET database. MIT's metric indicates 11.7% of the US labour market is technically replaceable by AI, with coding roles demonstrating significant overlap. This data serves as a critical constraint, enabling organisations to prioritise training and infrastructure investments, though high applicability does not guarantee job displacement.
Subscribe for Weekly Updates
Stay ahead with our weekly AI and tech briefings, delivered every Tuesday.
