Federal Reserve Bank of Dallas
RSS FeedJob postings show early signs of AI automation impact
Original Published: September 1, 2026
🎯 Sentiment Impact: Concerning
📋 Summary
- Two-thirds of Texas firms surveyed in May 2026 reported using generative AI, up from 40% two years prior, and a new Dallas Fed analysis of millions of job postings finds GenAI is measurably reshaping labor demand.
- Job openings for occupations with tasks automatable by GenAI fell 5% relative to less-exposed occupations by end of 2023 and roughly 8% by Q1 2025, using a task-exposure metric built on Anthropic's Claude usage data mapped to the O*NET occupation database.
- Firms whose pre-ChatGPT job mix was 10% more automatable posted 2 percentage points fewer automatable-task jobs afterward (a nearly 50% relative reduction), with effects concentrated in software development, web design, and other computer-heavy and white-collar roles.
💡 JR Insights
- 💼 Implications: Existing firms—not just AI-native startups—are driving the pullback in automatable job postings, reducing such postings by 8-9% by early 2026, showing this is a broad-based shift in hiring behavior, not just a new-firm phenomenon.
- 🚨 Risks: Because most online job ads require little prior experience, the automation-driven pullback disproportionately hits new labor market entrants, contributing to unusually high unemployment among recent college graduates.
- ✨ Advice: Students and recent graduates should factor occupational AI-exposure data into their field of study and job search strategy, since firms are actively shifting job postings away from highly automatable roles.