context: The PRC's AI governance currently focuses on content safety, data compliance and algorithmic ethics, but lacks systematic early warning for employment impacts. The 20th Central Committee's Fourth Plenum (October 2025) required improving employment impact assessment to address new technology development. The Ministry of Human Resources and Social Security announced in January 2026 it will issue documents on AI employment impacts, while the 2026 Two Sessions Government Work Report pledged measures adapting to AI development to promote employment and entrepreneurship.
AI application employment impact assessment mechanisms should be established, proposed Ma Yide 马一德 an NPC (National People's Congress) deputy and University of Chinese Academy of Sciences School of Intellectual Property Law dean during the 2026 Two Sessions.
Ma warned the current policy gap, if not promptly addressed, could evolve into major risks affecting social stability. He argued policy currently emphasises cultivating new occupations and expanding new employment forms to absorb AI-driven industrial changes, but should strengthen prevention of employment shocks.
Ma identified four major employment impacts from AI
- white-collar positions face systematic replacement, with translation, customer service and basic design positions experiencing significant demand contraction
- 'AI plus senior expert' model is replacing traditional 'senior expert plus junior assistant' structures
- around 300 million global jobs may be affected by AI automation, with considerable proportions concentrated in knowledge-based and clerical positions
- youth employment pressure intensifies as enterprises reduce entry-level hiring
- the 2026 graduate cohort will reach 12.7 million, while sectors with strongest employment absorption including internet, finance and education training face greatest AI impacts
- middle-aged workers confront 'skill depreciation'
- workplace core forces in their thirties and forties see professional skills underpinning livelihoods accelerate in devaluation, lacking young people's advantages in learning new technologies while not yet reaching retirement age, becoming vulnerable groups in technological transformation
- traditional re-employment pathways ineffectiveness
- previously displaced workers could transfer to service industries, but customer service robots, AI shopping guides and intelligent financial advisors proliferate, weakening service industry 'reservoir' functions
- flexible employment positions including food delivery, express delivery and ride-hailing also face potential replacement from unmanned delivery and autonomous driving
Ma proposed establishing employment impact assessment systems for major technology applications, referencing environmental impact assessment institutional approaches. Enterprises deploying large-scale AI human replacement should establish pre-assessment and process monitoring mechanisms. Employment effects should be incorporated into AI industrial policy consideration frameworks, with projects having negative net employment effects requiring enterprises to invest matching retraining and reemployment placement funds.
For technology-driven unemployment social security, Ma suggested MoHRSS (Ministry of Human Resources and Social Security) research revising unemployment insurance regulations, adding 'technology-driven unemployment' identification standards and special relief measures. For recognised technology-driven unemployed workers, appropriately extend unemployment insurance benefit periods, raise training subsidy standards and establish 'training period living subsidy' systems, resolving workers' dilemmas between supporting families and full-time training.
Regarding financing, Ma proposed exploring multi-source financing mechanisms for technology-driven unemployment insurance funds. This could include researching moderately adjusting social insurance contribution ratios for enterprises with AI applications reaching certain scales or establishing 'technological progress adjustment funds' requiring AI application beneficiary enterprises to share partial social costs, embodying 'whoever benefits bears responsibility' fairness principles
Yao Jinbo 姚劲波 an NPC deputy and 58.com CEO also proposed implementing 'national AI skills and literacy enhancement action', establishing national AI talent cultivation special funds and building employment monitoring systems led by MoHRSS departments jointly with universities, research institutions and recruitment platforms.