AI adoption boosts productivity in Chinese enterprises, but skills and job concerns persist: ILO

Artificial intelligence adoption is delivering significant productivity gains across Chinese enterprises, but concerns around job displacement, skills gaps and future incomes are emerging alongside the technology’s expansion, according to new research by the International Labour Organization (ILO).

The ILO research brief draws insights from 21 enterprises across manufacturing, finance, business services, construction, education, media and travel, alongside a survey of 1,591 professionals. 

The study found that all 21 companies interviewed were either already using AI or had concrete plans to adopt it. While the firms varied significantly in size and sector, AI adoption was primarily being used to improve efficiency, enhance quality and control costs.

AI delivers measurable productivity gains

Companies that track AI’s impact reported substantial improvements across several business functions.

One insurance company saw 300 customer-service employees increase their daily handling capacity from 6,000 to 15,000 customer issues after introducing AI, representing a 150 per cent increase. Another large insurance group reduced recruitment cycle times from 30 days to 13 days, while a smart manufacturing facility reported a 30 per cent improvement in production efficiency.

Other companies reported efficiency improvements of between 30 and 100 per cent across selected workflows, while one travel services company estimated that AI-enabled data handling provided labour capacity equivalent to five to six full-time employees.  

However, the ILO cautioned that these productivity figures are based on companies’ own reporting and have not been independently verified. It also noted that most enterprises lack systematic frameworks for measuring AI’s broader impact.

This measurement gap extends beyond productivity. The research found limited assessment of AI’s effects on job quality, working conditions and how workers’ tasks are changing.

AI adoption is largely centred on hybrid workflows

Rather than fully replacing employees, the enterprises studied are predominantly integrating AI into existing workflows alongside human workers.

The study found that AI is particularly being applied to repetitive and data-intensive activities, including document processing, customer queries, CV screening, data collection, knowledge management and other administrative tasks.

The research identified three broad organisational approaches to AI adoption: centralised specialist teams, business-embedded AI integration and bottom-up adoption driven by employees. 

Most companies remain at the stage of using AI to improve existing tasks or optimise established processes, while a smaller number are moving towards AI-enabled changes to their business models.

This hybrid model means employees continue to play a role in directing, reviewing and validating AI-generated outputs, particularly where accuracy and domain expertise are important.

Skills gaps emerge as a key barrier

While AI adoption is widespread, companies reported challenges around employee skills, resistance, output quality, data security, regulation and integration with existing systems.

The research found evidence of age-related differences in AI capabilities in some companies, with one education technology firm reporting weaker AI application capabilities among employees over 40.

The study also identified a gap between technical AI expertise and industry-specific knowledge. Some companies reported that AI specialists lacked sufficient understanding of business applications, while industry experts often lacked the AI skills needed to use the technology effectively.

The findings suggest that AI adoption will require more than technical training. Workers will also need the ability to identify suitable use cases, interact effectively with AI tools and critically assess their outputs.

Workers remain divided over AI’s impact on jobs and income

The survey found that 56 per cent professionals viewed AI adoption as an inevitable trend, while 47 per cent believed AI would create more jobs than it displaces.

At the same time, 39 per cent expected AI adoption to result in lower incomes. Around one-third of respondents also expressed concerns about AI’s implications for social stability and human uniqueness. 

The ILO research suggests that displacement pressures could be concentrated in routine clerical, administrative and customer-service roles because these occupations contain a higher proportion of repetitive and data-intensive tasks.

At the same time, the researchers found that AI adoption is also creating opportunities for workers to move towards higher-value activities where human judgement, expertise and oversight remain important.

ILO highlights lifelong learning and transition support

The research identifies lifelong learning, worker transition support, better AI impact measurement and wider access to AI for smaller businesses as key areas requiring attention.

The ILO said skills programmes should particularly support mid-career and older workers who may face greater barriers to adapting to new AI-enabled ways of working. It also highlighted the need to help workers transition from tasks affected by automation towards higher-value activities.

For smaller businesses, the research points to shared platforms, training programmes and affordable AI services as potential ways to broaden access to the technology.

The study also calls for more comprehensive frameworks to measure AI’s effects, covering not only productivity but also job quality, working conditions, task composition and how the benefits of technology are distributed.

The findings reinforce the ILO’s broader view that AI can either automate or complement human labour, depending on how the technology is integrated into work and the extent to which workers remain involved in overseeing and performing tasks.

The key takeaway is that AI strategy and people strategy cannot be treated separately. The research suggests that the value of AI will depend not only on how quickly organisations deploy the technology, but also on how effectively they redesign work, build the skills needed to use AI and support employees through the transition.