Goldman Sachs says AI is reshaping junior jobs, not just eliminating them

AI is not just taking jobs at Goldman Sachs and in the broader labor market. It is splitting work into two camps: tasks it can replace and tasks it can improve, and that distinction matters most for analysts and associates whose careers still depend on the shape of junior work.

Goldman Sachs Research laid out that framework in April 2026, arguing that substitution and augmentation are moving at the same time. Its economists estimated that AI had trimmed monthly payroll growth by roughly 16,000 jobs over the past year and nudged unemployment up by about 0.1 percentage point. At the same time, the firm said augmentation had added about 9,000 jobs per month in some sectors, a reminder that AI can hollow out some roles while strengthening others.

For Goldman employees, the practical lesson is less about headline job loss than about where the first redesigns will hit. The report said the negative effects are landing disproportionately on younger, less-experienced workers, which points straight at the analyst and associate pipeline. AI can take over first-pass research, document review and summarization, the kind of repetitive work that has long served as training ground on Wall Street. That does not mean junior jobs vanish overnight, but it does mean the old apprenticeship model is likely to compress faster than many people expect.

That shift has real consequences for career development inside Goldman. If machines handle the first mile of the work, junior bankers may spend more time on judgment, client interaction and issue spotting, the skills that still separate a competent analyst from someone ready for a promotion track. The pressure on managers is just as obvious: headcount planning can no longer assume that every rung of the pyramid needs the same volume of rote assignments to justify itself. Firms that keep staffing models frozen while the work changes will end up with too many people doing lower-value tasks, or too few people learning the work that still requires human judgment.

Goldman’s researchers tried to capture that nuance by combining AI displacement scores with an IMF-style complementarity index. In plain terms, they were asking not just which jobs are exposed, but which ones become more valuable when AI is added. That is the distinction employees should care about most. The winning career strategy is not to treat AI as either a job-killer or a productivity miracle, but to build the skills that are hardest to automate, use AI for the first draft, and reserve human effort for decisions, interpretation and client-ready advice.