无畏的未来：普华永道2025年全球AI就业晴雨表

AI makes people more valuable Key Findings Our analysis of close to a billion job ads and thousands of company financialreports across six continents suggests: 3xHigher growth in AI is making workers significantly more productiveand creating value for companies revenue per workerIn industries more exposed •Industries most able to use AI have 3x higher growth in revenue generated byeach employee 100% •Since 2022 when awareness of AI’s power surged, productivity growth inindustries best positioned to adopt AI has nearly quadrupled (while falling Of industries areincreasing AI usage •Workers with AI skills like prompt engineering command a 56% wage premium(up from 25% last year), suggesting the value these workers bring Including industries lessobviously exposed toAI such as mining and •100% of industries are expanding their usage of AI (even industries lessobviously exposed to AI such as mining and construction), indicating that 66% Companies are using AI to make workers morevaluable and productive, not just to streamline Faster skill changein AI-exposed jobs Up from 25% last year.Change is fastest in •Contrary to fears about job losses, job numbers – and wages – are growingin virtually every AI-exposed occupation, including the most highly AI is creating deep change in the skills workersneed to succeed •Skills sought by employers are changing 66% faster in occupationsmost exposed to AI (like financial analyst) versus least exposed (like physical Five Implications forBusiness Leaders Our data suggests businesses are starting to see benefits from AI measurable in revenue peremployee. To compete, business leaders should have a plan to capture that opportunity. Manyorganisations are starting to use AI for isolated use cases. But the real benefit comes when AI isused to transform value creation at an enterprise-wide level, generating new revenue streams 2 Our data makes clear that companies are using AI not just to control headcount but rather to helpworkers create more value. Companies who use AI only to reduce staff numbers may miss out on themuch bigger opportunities to use AI to claim new markets or generate new revenue streams. 3 With AI agents at their command, workers can achieve much more. Business leaders who adoptagents early won’t just cut costs – they can create organisations that think, adapt, and execute fasterthan competitors. Get the greatest value from your agents by enabling them to work as a team –sharing context, operating across platforms, and learning from one another – with PwC’s agent OS. 4 As AI creates huge churn in the skills workers need, build a clear, data-based picture of skills gapsand create a plan for closing them. 5 Unlock AI’s transformative potential by building trust.Our research suggests the growth dividend from AI is not guaranteed and depends on more than just technical success – it also hinges on responsible deployment, clear governance and public andorganisational trust. AI can make workersmore valuable, not less –even in the most highlyautomatable jobs 01 This is our conclusion from analysing close to a billion job adsand thousands of company financial reports from six continentsto uncover AI’s impact on jobs, wages, skills, and workers’productivity. Our analysis captures AI’s impact as Agentic AI – We analysed AI’s impact on both augmentable jobs (jobs that contain manytasks in which AI can enhance or support human judgment and expertise),and automatable jobs (jobs that contain many tasks that can be autonomously AI-exposed jobsare jobs that contain many tasks in which AI can be used.Example jobs: financial analysts, data entry workers.1We use ‘AI-powered’ as Augmentable jobsare AI-exposed jobs in which AI enhances orsupports human judgment and expertise on many tasks. Example jobs: Automatable jobsare AI-exposed jobs in which AI can carry out many tasks.Example jobs: software coders, customer service workers. AI-exposed industriesare industries that contain many job roles in whichAI can be used. For example, software publishing is an AI-exposed industry, We analyse AI exposure (the ability to use AI) as a proxy for AI uptake,though actual levels of AI uptake may lag AI exposure. As levels of AI usagegrow, the effects we find in this report may be magnified. Let’s meet two workers whose experiences show what the future holds for millionsof workers, our data suggests. First, let’s meet Amina. She is an informationanalyst, and AI is augmenting her analytical abilities. Amina now instructs AIagents to do research and draft reports for her, freeing her to spend more timeinterpreting and refining the reports, servicing clients, and developing new Next, let’s meet John. John is a customer support agent, and AI is automatingmany parts of his role including handling the simpler queries John dealt with inthe past. John worried that AI might replace him, but in fact AI has enhanced hiscapabilities. Now John helps customers with more compl