Overview

• Amodei predicts that advancing AI could erase roughly 50% of entry-level white-collar roles over the next one to five years, raising US unemployment to between 10% and 20%.
• Anthropic this month rolled out its latest Claude Opus4 and Sonnet4 models, which it says set new standards for coding and advanced reasoning but prompted safety tests revealing potential misuse.
• He calls on AI companies and government to stop downplaying job risks and to develop honest communication strategies with workers and legislators.
• Recommended mitigation measures include upskilling employees in AI tools, boosting public awareness, considering a ‘token tax’ on AI usage, and crafting regulatory frameworks focused on augmentation over automation.
• Economic researchers caution that labour-market adaptation and offsetting productivity gains may moderate AI-driven job losses, highlighting uncertainty around the scale of displacement.