AI对就业的深层影响：算法共谋或成薪资杀手
最近，国际劳工组织研究部门的宏观经济政策专家埃克哈德·恩斯特(Ekkehard Ernst)在北京接受南华早报访问。
当讨论到人工智能（AI）对就业市场的作用时，他指出，相比“机器人夺走工作岗位”的恐慌说法，更值得警惕的是“算法合谋”现象，这种趋势可能在无形中拉低薪酬水平并危害职场环境。
机器人真的会抢饭碗吗？
公众普遍担忧人工智能可能导致大量人员失业，然而这类破坏性后果实际上被夸大了。“我认为目前尚未达到AI显著冲击就业市场的程度。”
美国Anthropic人工智能公司本月公布的研究揭示出一个显著的“落地差距”。该报告指出，尽管人工智能具备执行多种高端工作的潜力，但在实践中推进缓慢，原因涉及法规障碍、系统整合难度及对人力监督的依赖等因素。
人工智能确实在某些行业中发挥作用——特别是在软件开发和初级职位方面，但将其与青年人就业难题泛泛而谈并不科学。
中国的年轻群体失业情况（16至24岁为16.1%，25至29岁为7.2%）与一些欧洲国家（可能超过20%）相比，并不算异常突出。
青年人求职困境，“主要源于当前经济增长放缓，而非人工智能这个单一变量”。
埃克哈德·恩斯特
“根据我们对中国人口结构演变的观察，真正的问题可能是劳动力不足，而不是岗位稀缺。”
人工智能推动工作变革与升级
在就业环境中，人工智能更多体现为重新定义和优化职业角色，而非全面取代人力，因为它使员工能专注于更具挑战性和高价值的工作内容。
恩斯特驳斥了“总量恒定误区”——即认为就业需求固定不变的看法。他用车辆制造和医疗服务举例，效率提高导致成本降低后，消费需求随之增长，进而刺激了用工需求。
“基于我们观测到的人口走势，中国面临的风险是缺少工人而非缺乏职位。”在诸如中国这样步入老龄化的社会里，人工智能可在老年照护和健康服务领域填补重要空白。
“并非没有工作了，而是工作形式变了。”
伴随科技重塑就业格局，个体将面对持续学习与能力更新的压力，这需要充足的时间和公共资源支持，以便顺利过渡到新型职务。
“算法合谋”与数字化劳动监管
他强调，最令人忧虑的是人工智能借助“算法合谋”干扰薪酬设定流程，这或将增强劳动力市场的垄断倾向，使得雇主以非市场竞争手段压制薪资。
随着大型企业引入人工智能进行人力资源管控，这些算法或许会被用于解析对手信息，最终导致提供的报酬低于公平竞争条件下的应有水准。
新经济形态的发展令就业状况更加多元。越来越多刚毕业的学生加入灵活就业行列，因此政府亟需出台调控策略，防止陷入“逐底竞争”——即片面追求极致效能却漠视劳动者的基本承载力与安全保障。
他提议可参照金融业的管理模式，在算法上线之前实行必要的“抗压检验”。
“尽管我们已设立面向工厂职工的职业安全规范，但政府部门也应当构建专门机构，负责监督虚拟办公空间。”
AI ‘collusion’ forcing down wages a bigger threat than job-stealing robots: ILO economist
The threat to employment posed by artificial intelligence was not a “robot apocalypse” that would steal jobs, but “algorithmic collusion” that could quietly erode wages and workplace safety, Ekkehard Ernst, the International Labour Organization’s chief macroeconomist, warned in Beijing on Tuesday.
While public anxiety frequently centred on the potential for AI to trigger a mass wave of unemployment, Ernst said its disruptive potential had been overestimated.
“I don’t think that we are anywhere close to major disruption of labour markets,” he said.
Citing a study released by American AI company Anthropic this month, Ernst noted a stark “implementation gap”. The study showed that while AI was theoretically capable of performing many high-paying tasks, real-world adoption lagged significantly due to regulatory hurdles, system integration complexities and the need for human oversight.
While AI was having an impact on specific sectors – notably software engineering – and entry-level roles, Ernst said broader concerns about its impact on youth employment were misplaced.
Comparing China’s youth jobless rates – 16.1 per cent for 16- to 24-year-olds and 7.2 per cent for 25- to 29-year-olds – to those of some European countries where the figure could be over 20 per cent, he said they were not exceptionally high.
Instead, the struggle for young people was “mostly related to the current economic slowdown, more than to specific AI”.
Ekkehard Ernst
“Considering the demographic developments that we see in China, we are risking running out of workers, not running out of jobs.”
In the labour market, AI was more about reshaping and transforming jobs rather than massively replacing them, he said, because it allowed workers to focus on complex, high-value-added tasks.
Ernst dismissed the “fixed-pie fallacy” – the idea that labour demand is finite. Using the automotive and healthcare sectors as examples, he said that as increased efficiency lowered prices, market demand rose, increasing the need for labour.
“Considering the demographic developments that we see in China, we are risking running out of workers, not running out of jobs,” Ernst said, adding that in ageing societies like China, AI could bridge critical gaps in elderly and medical care. “So the job has not disappeared. It’s just different.”
But he also acknowledged that people faced the challenge of lifelong learning and reskilling as the labour market was reshaped by technology – a process that required time and public investment to prepare workers for new roles.
He said his primary concern was the way AI could distort wage-setting processes through “algorithmic collusion”, which could exacerbate labour market concentration and allow employers to suppress wages non-competitively.
Ernst said that as major corporations adopted AI-powered tools for human resources management, algorithms could be used to analyse competitors’ data, leading to lower wages being offered than would be the case in a truly competitive market.
The rise of the “platform economy” also complicated the employment landscape, he said. With more young graduates taking on gig work, Ernst said governments needed to intervene to prevent a “race to the bottom” that would maximise efficiency while remaining indifferent to the basic endurance and safety of human workers.
He proposed a regulatory model similar to that in the financial sector, with mandatory “stress testing” for algorithms before they were deployed.
Ernst said that while “we have occupational safety regulations for factory workers”, there was also a need for governments to create agencies to oversee the digital shop floor.
向右滑动，查看英文版
本采访以“AI ‘collusion’ forcing down wages a bigger threat than job-stealing robots: ILO economist”为题，发表于南华早报（South China Morning Post），作者：Carol Yang。本次采访所表达的观点仅代表恩斯特先生个人观点，不一定代表国际劳工组织的立场和观点。