AI's impact on jobs: Automation vs. Augmentation

Analysis from Stanford University indicates a 2.7% decrease in employment for individuals aged 22 to 25 since the widespread adoption of ChatGPT. This effect is more pronounced in sectors highly exposed to AI, such as finance, software development, and creative industries, where the employment hit has reached 12.8%. The OECD has also observed differences in job posting trends between highly AI-exposed sectors and less exposed ones, noting that the UK, with its concentrated service sector, appears particularly vulnerable to potential AI-driven job losses.

AI models are rapidly advancing, with current large language models (LLMs) now capable of completing complex tasks that previously took hours. Some of the latest models are even approaching the ability to self-develop software, a pattern also emerging in financial analysis and legal work. The use of AI agents for automated tasks has surged, consuming trillions of tokens. However, the associated costs have become so high that many companies are now rationing AI usage, suggesting potential limits to automation depending on the task's expense.

In response to these high costs, many companies, including Western firms, are reportedly shifting towards cheaper AI alternatives derived from Chinese models that are often provided freely.