GCRI · Artificial Intelligence and Autonomous Systems
Labour and social effects
AI changes which tasks people do, how work is organised, which skills are valued and how information spreads. Effects differ by occupation, country, gender and age, and depend on how firms and governments choose to deploy it. This page sets out how AI affects work and society, how exposure and impacts are measured, and what evidence serves the labour ministries, employers, unions and educators involved.
Labour and social effects on the record.
Labour and social effects: how the risk develops.
Work, education, public trust and inclusion. The dossier below sets out how the risk develops, what is measured, where the data come from, how it is modelled and who decides.
Tasks change before jobs do
AI automates or assists specific tasks within occupations. Jobs are reorganised around remaining tasks before they disappear or new ones form.
Exposure is uneven
Clerical and some professional tasks are highly exposed to generative AI, while manual and interpersonal work is less exposed. Women are over-represented in some highly exposed clerical jobs.
Skills and training lag
Workers need new skills to use AI well, and those displaced need retraining. Training systems often respond more slowly than technology adoption.
Information ecosystems shift
Generated content, recommendation systems and synthetic media change how people find and judge information, with effects on trust and public debate.
How labour and social effects is modelled, measured and tested.
The model, its assumptions, its data and its uncertainty, in the form a scientist or engineer would check. Structures follow published methods. The values drawn are illustrative, not findings.
System model
The couplings between the systems in labour and social effects: signs of influence, a reinforcing and a balancing loop, and cross-links where stress jumps.
AI’s effects on work vary across domains.
- HealthCouplingClinical and administrative tasks are exposed differently.What to watchTask exposure by occupation
- FoodCouplingAgricultural work is less exposed to generative AI than clerical work.What to watchExposure by sector
- EnergyCouplingGrid operations need new digital skills.What to watchSkills gaps
Labour studies are designed for the Exponential Technology series. The ILO’s 2023 working paper estimates exposure, not job loss, and finds clerical work most exposed.
- Firm-level adoption studies.
- Worker surveys on task change.
What is measured, and how well.
Each indicator carries its unit, its cadence, the kind of source behind it and the evidence label a value would carry. No values are shown: a programme supplies them under its own authority.
| Indicator | Unit | Cadence | Source type | Label |
|---|---|---|---|---|
| Employment in occupations highly exposed to AI | percent of employment | per assessment | Exposure studies using occupational data | ESTIMATED |
| Firms using AI | percent of firms | annual | Business surveys | REPORTED |
| Workers using AI tools at work | percent of workers | per survey | Labour force and worker surveys | REPORTED |
| AI-related job postings | share of postings | monthly | Online job posting data | OBSERVED |
| Participation in AI skills training | percent of workforce | annual | Training surveys | REPORTED |
| Wage changes in exposed occupations | percent change | annual | Wage statistics | Published statistics |
Where the evidence comes from.
Named public datasets and institutional records, each with its licence class. Sources by agreement are used only under their terms and never republished.
| Source | What it gives | Licence class |
|---|---|---|
| ILO studies on generative AI exposure of occupations | Occupational exposure estimates by country and gender | Public, ILO terms |
| OECD Employment Outlook and AI in work studies | Evidence on AI adoption and labour effects | Public, OECD terms |
| Eurostat ICT usage in enterprises survey | Share of enterprises using AI | Open, Eurostat reuse policy |
| Online job posting data such as Lightcast summaries | Demand for AI skills | Commercial; public summaries (to check) |
| Stanford AI Index | Compiled indicators on AI adoption, investment and labour | Public, licence to check |
Who decides, and what the evidence serves.
Artificial Intelligence and Autonomous Systems prepares evidence. The decisions below are taken by the authorities named, under their own rules, and never by the platform.
| Who decides | The question | Evidence this page serves |
|---|---|---|
| Labour ministries | Which workers and sectors need transition support | Exposure analyses; adoption tracking |
| Employers and unions | How to introduce AI at work and share its gains | Adoption and wage evidence |
| Education and training bodies | Which skills to teach and to whom | Skills demand evidence |
Outputs this topic produces
- Occupational exposure analyses by sector, gender and region
- Adoption tracking from business and worker surveys
- Evaluation designs for AI training programmes
- Evidence notes for social dialogue on AI at work
What researchers ask about social effects.
- Which tasks can current AI perform well, and how does that map onto occupations?
- Does AI substitute for workers, complement them or change what their jobs consist of?
- How does AI change learning and assessment in education?
- How do errors, opacity and misuse affect trust, and who is excluded?
Tasks, not jobs.
Jobs are bundles of tasks. AI tends to automate or assist specific tasks, such as drafting, summarising, coding or classifying images, which changes the content of jobs rather than simply removing them. Whether that raises productivity and wages or displaces workers depends on how organisations redesign work and how gains are shared.
In education, AI can personalise support but also undermine assessment and the learning that comes from effort. Trust depends on experience: visible errors, unexplained decisions and synthetic content erode it, while accountable use can build it. People with limited digital access, skills or language support risk exclusion.
Evidence on social effects.
- 01
Task exposure analysis
Mapping AI capabilities onto occupational task data. Exposure is not the same as displacement.
- 02
Field and natural experiments
Studying workplaces and classrooms that adopt AI, using social and behavioural science. Early studies may not generalise.
- 03
Surveys of trust and use
Measuring how people use and perceive AI. Attitudes change quickly and vary by context.
- 04
Capability evaluation
Linking what systems can actually do, from AI evaluation and assurance engineering, to claims about social effects.
How it is studied, and what the methods cannot say.
- 01Social and behavioural scienceHuman oversight studies · Affected community research · Workplace adoption studiesApplied here
- 02Probability and statisticsUncertainty on benchmark scores · Calibration assessment · Audit samplingApplied here
- 03Artificial intelligence and machine learningDistribution shift analysis · Fine-tuning for a specific use · Interpretability methodsApplied here
- 04Robotics and automationScenario-based testing · Collaborative robot safety · Operating domain monitoringApplied here
Where this page’s references lead.
The bodies and documents this page is read against. Status and editions change; each link goes to the issuing body.
Questions and answers.
Will AI take jobs?
What is occupational exposure to AI?
Who is most exposed to generative AI?
Will AI cause mass unemployment?
Which workers are most exposed?
How does this topic relate to Future of Work and Skills?
Does the platform advise employers?
Contribute evidence from this page.
Reports is the Consortium’s quarterly publication. What you know about this subject can become evidence, a case or a correction in the next edition, credited to you and published under CC BY 4.0 where it is your own work.
- EmailGateway GW-GCRISend it from this pageEvidence, a correction, a case, a method, a review, a translation or profile input. It arrives tagged with this page, so the editor who owns it receives it.
- ZenodoOpen depositDeposit it and submit to the communityPublish your own record on Zenodo with your ORCID iD, then submit it to the Reports community there. Curators review it; the DOI stays yours.
- ReportsGCRISee every way to contributeThe contribution types, how each is reviewed, what credit you receive, the licence and what is never accepted.
Never send personal data, sensitive infrastructure detail or anything you have no right to share. A contribution is reviewed before use; sending it confers no authorship or editorial rights, and sponsorship never buys a finding.
Bring a method, a dataset or a question.
Information and enquiries are open today. An enquiry receives a reference and names the next step; applications, payments, compute allocations and registrations open only when their team, terms and review capacity exist.
Legal documents.
Maps, data and figures on this site are shown for context and general information. They are not advice, forecasts or warnings; the limits of the information are set out in the Disclaimers and Data Notices.