Why AI Hasn't Taken Your Job Yet

Remember when AI was supposed to snatch jobs from radiologists, software engineers, and pilots? It hasn't happened. Despite predictions, these jobs are still here. The reason is simple: AI can't handle the messy, human parts of these roles.

The Limits of AI

Benjamin Todd, president of 80,000 Hours, explains why AI hasn't wiped out certain professions. Sure, AI can automate parts of a job, but not the entire role. Ask the workers, not the executives. Radiologists spend only a third of their time interpreting images. The rest involves tasks that AI can't do, like consulting with doctors and managing patients. This pattern repeats across various industries. Even software engineers, whose field AI has disrupted the most, are still in demand. Job openings for software engineers at tech firms hit over 67,000 in 2026, doubling since mid-2023. Why? Because AI tools have boosted productivity, not replaced humans.

Focus on 'Safe Skills'

So what's the real risk? It's when AI can handle enough of a role to make workers redundant. That's when wage pressure kicks in. Todd suggests focusing on 'safe skills' instead of 'safe jobs.' Skills that are hard for AI, complement AI, produce highly desired outputs, and are tough for others to master are the key. Think strategy, long-term research, and social coordination. But here's the kicker: the balance could shift fast if AI systems start handling bigger projects.

Automation isn't neutral. It has winners and losers. While some jobs remain secure, others may not be so lucky. So, what's the future? Should we fear displacement or embrace retraining? The jobs numbers tell one story. The paychecks tell another.