The rapid emergence of AI increases the need to understand its social and economic adoption. A national survey shows extensive use both at home and at work, with greater adoption among people with higher education and concern about job disruption. Business adoption is common in large and small companies, with demographic differences in its use.
The Latam-GPT team is refining its model with Spanish materials from Latin America, coordinated by Chile’s National Center for Artificial Intelligence, to create local artificial intelligence that reflects the region’s languages, cultures and histories. Despite obstacles such as talent flight, it aims to be a public infrastructure for educational applications and public services, highlighting institutional diversity and cross-border collaboration.
The paper provides a framework to understand how supply shocks affect inflation, identifying eight types of shocks, six transmission channels and seven amplification mechanisms. Some supply shocks and economic characteristics are more prone to generating inflation, and amplification can be significant with multiple mechanisms involved.
Public debt levels in advanced economies have decreased since the pandemic but remain high. Deficits in 2024 were elevated in most countries and fiscal effort will be required to stabilise debt. In the EU, public debt is projected to increase in the medium term, and although the primary balances required to stabilise debt are feasible, deficit cuts are significant in several countries. Adjustments are complicated due to spending pressures, population ageing and political resistance. Some countries in need of adjustment may be vulnerable to market sentiment changes.
Self-construction benchmarks are essential to evaluate the capabilities of Large Language Models (LLM). We use agentive artificial intelligence to have LLM generate and evaluate practical exams in Finance, Business Operations, Management, Computing, and Mathematics. Although leading models achieve median scores of 65-79%, they show weaknesses in data manipulation and financial calculations. LLM-generated benchmarks can offer a cost-effective, scalable, and updatable way to measure the job capabilities of artificial intelligence.
