MOHRSS Promotes 'AI Adaptation Employment Action': What Does It Mean for Ordinary Workers?

China's Ministry of Human Resources and Social Security is promoting an 'AI Adaptation Employment Action' aimed at adapting to AI development and promoting employment. The policy acknowledges that AI's employment impact is already happening, not a future possibility. The wording 'actively respond' signals a moderate stance, choosing adaptation over confrontation. The policy also calls for assessing the employment impact of major policies, projects and productive-capacity layouts, meaning AI industry investment may be recalibrated to include job effects. The most concrete measure is the 'AI technology skills for all' action, covering basic AI education for all workers, digital engineer training for specialists, and AI training for students. Targets for the 15th Five-Year Plan period include 10 million subsidized vocational training participations per year and 40 million new skill certificates over five years. Labor market data supports this shift. In Q1 2026, recruitment demand for editors fell 29% year on year, customer service 23%, and graphic design 21%. Demand for large language modeling skills rose 73%, AI product managers 81%, and AI engineers and trainers 17% each. One study found that over a quarter of new online recruitment pool positions have high AI substitution exposure, while about 21 million jobs were created through task restructuring. AI is reorganizing tasks rather than simply eliminating occupations. Those who master AI are benefiting. AI-related recruitment on new-economy platforms rose 789.47% year on year in January to July 2026, with new AI positions averaging 63,160 yuan per month. Those performing standardized tasks face pressure: one data department of 50 people was dissolved, and graphic designers saw pay drop. Entry-level roles are shrinking most sharply, with demand for machine learning roles requiring under one year of experience down 71.43%, image algorithms down 66.67%, and software testing down 52.94%. For individuals, this means stability now depends on adapting to changing job content, training has a shorter shelf life, and multi-skilled, closed-loop capability is becoming the baseline. Policy can offer transition opportunities, but what decides success is whether workers learn skills that seem unnecessary while they can still cope.