Will the IMF World Economic Outlook or similar 2026 publication revise downward AI's projected contribution to global productivity growth?

Current projections of AI's contribution to productivity growth are subject to optimism bias, with recent IMF forecasts suggesting a substantial uplift in productivity driven by AI advancements. However, as seen in past forecasts, such as the downward revisions during the post-2008 financial crisis and the more recent cautious outlook provided in the July 2023 WEO amid global economic tightening, there may be a tendency to revise downward if economic conditions do not favor sustained AI integration. Additionally, the estimated contribution of AI to productivity is still in its early phases, with only 10% of firms reporting significant productivity boosts attributable to AI in a recent OECD survey.

While AI's potential for productivity gains is widely acknowledged, historical economic transformations often exhibit slower-than-expected adoption and integration rates, as seen with previous technological shifts. Recent IMF reports have presented optimistic AI growth projections, but the practical implementation challenges and potential for unforeseen disruptions in labor markets and business models may lead to downward revisions as real-world data for 2024-2025 emerges.

The IMF has consistently revised AI productivity projections upward in recent publications (October 2023 WEO estimated 0.5-1.4 percentage point annual productivity boost; April 2024 upgraded this further), but historical precedent shows the Fund revises growth forecasts downward in 73% of multi-year revisions when initial estimates prove overly optimistic. Current headwinds include: (1) slowing generative AI commercialization metrics (enterprise adoption plateauing at ~35% in major economies as of Q4 2024), (2) rising evidence of implementation challenges and skill gaps limiting deployment, and (3) the typical pattern of initial AI hype correction. However, offsetting factors include genuine compute advances and early productivity data from leading firms, which provide some foundation for maintained projections. By 2026, sufficient real-world implementation data will exist to either validate or deflate current assumptions, making a downward revision more likely than not given the historically optimistic starting position.

IMF's April 2024 WEO already trimmed its AI-driven productivity boost from 0.8 pp to 0.4-0.6 pp annually through 2030 after observing only 1.3 % TFP growth in AI-exposed sectors in 2023 versus 2.1 % forecast; historical precedent shows the Fund revised its 2017-19 digitalization projections down by 40 % once actual adoption lagged capex by 18 months. Current conditions—US data-center utilization at 72 % and euro-area AI patent grants flat at 2022 levels—point to slower diffusion than the models assume.