Will AI replace my job, and what should workers check?
No exposure score can tell you whether a particular job will disappear. Generative AI may change some tasks, and the result depends on how employers deploy it, how work is reorganized, and whether workers have a say.
Does task exposure predict job loss?
No. The International Labour Organization says exposure indicators are signals of possible task change, not forecasts of employment outcomes. Its 2025 global index found clerical occupations had the highest exposure, while many other jobs combine tasks that models may handle and tasks that still need people. ILO guide to exposure measures ↗ · ILO 2025 working paper ↗.
What does a worker-protection review measure?
Look for evidence about wages, safety, hours, contractor conditions, grievance processes, and worker voice in the AI supply chain. Those are questions about the people who build, label, moderate, and operate AI. They are separate from whether a company expects AI to reduce headcount or change job tasks. The catalog’s worker findings and sources are on compare models and compare providers.
When should I use a non-AI route?
Try the source document, a calculator, spreadsheet, library, or government reference when the task is routine or needs exact records. If AI helps with a draft, give it one bounded task, check the result against the original source, and keep a person responsible for the final decision.
How can a workplace make adoption accountable?
Ask which tasks change, who reviews errors, what happens to workload and pay, and how workers can challenge a harmful process. Measure the effect on the whole job and working conditions, not just how many tasks a model can perform. A worker-treatment score is evidence about documented safeguards; it is not a forecast of job displacement.