Anthropic's latest research highlights a significant gap between AI's potential and its current adoption in white-collar sectors such as business, finance, law, and computer science. The study reveals that while AI models like Claude theoretically cover up to 94% of tasks in computer and mathematics roles, actual usage is only 33%. This discrepancy is measured using a new metric called "observed exposure." The research indicates that highly educated, high-income female white-collar workers are most exposed to AI, with a 16-percentage-point higher proportion of women and 47% higher average income compared to less exposed groups. As AI capabilities advance, researchers warn of a potential "great recession for white-collar workers," akin to the 2007–2009 financial crisis, with unemployment risks looming. Currently, the impact is seen in reduced hiring, with job search rates for vulnerable professions dropping by 14% since the ChatGPT era, and employment rates among young workers aged 22–25 declining by 16%.