New study reveals early impact of AI on job market in UK (www.kcl.ac.uk)

A new King’s College London study by Dr Bouke Klein Teeselink analyzed millions of job adverts and LinkedIn profiles from 2021–2025 and finds that the public release of ChatGPT in November 2022 marked a clear inflection point in the UK labour market. Firms and occupations judged highly exposed to LLM capabilities cut total employment (−4.5% on average), with the effect concentrated in junior roles (−5.8%), pushing average pay up by ~£1,300 as workforce composition skewed toward senior staff. Hiring intentions at exposed firms dropped sharply (−16.3 percentage points). At the occupation level, highly exposed roles saw a 23.4% fall in job postings and a £2,951 (6.3%) decline in advertised salaries; technical jobs such as software engineers and data analysts experienced the steepest pullback, while customer-facing roles (e.g., sales) ticked up and remote work became more common. For the AI/ML community the study is an early empirical signal that LLM adoption is already reshaping demand: disruption is concentrated in higher-wage, technical segments (high-paying firms −9.6% employment; high-salary occupations −34.2% job postings), rather than across the board. Methodologically the paper leverages differential exposure to LLM-susceptible tasks to isolate effects, suggesting that displacement is uneven and task-specific. Key implications include a loss of entry‑level training pathways that could impair long-term skill pipelines, and a need for policy and retraining programs focused on interpersonal, customer-facing, and complementary skills that LLMs struggle to replicate.