EXPOSURE TO AI
10%
RESILIENT
OBSERVED IN REAL USE · Anthropic 2026
0%
of this role’s work is already showing up in real Claude usage (Anthropic Economic Index).
The last desk the machine reaches. Breathe.
10% of this role’s O*NET tasks are within reach of today’s AI. That is the core-weighted exposure score from Eloundou et al. 2023 (“GPTs are GPTs”). It measures a capability ceiling, not a headcount forecast. Certified hard to automate. Today’s AI barely touches the core of this one.
WHAT AI CAN ALREADY DO
- Direct items according to established routing schemes, using computer-controlled keyboards or voice-recognition equipment
- Search directories to find correct addresses for redirected mail
WHAT IT STILL CAN’T
- Clear jams in sorting equipment
- Operate various types of equipment, such as computer scanning equipment, addressographs, mimeographs, optical character readers, and bar-code sorters
- Sort odd-sized mail by hand, sort mail that other workers have been unable to sort, and segregate items requiring special handling
- Bundle, label, and route sorted mail to designated areas, depending on destinations and according to established procedures and deadlines
- "Move containers of mail, using equipment, such as forklifts and automated ""trains""."
THE HONEST PART. A percentage is not a pink slip. High exposure usually means a role shrinks and shifts toward judgment, direction and responsibility: the parts a model can’t sign its name to. Exposure ≠ displacement. Breathe.
"My job is 10% cooked. What’s yours?"
SOURCES: O*NET 30.3 occupational tasks · Eloundou et al. 2023 (“GPTs are GPTs”,
arXiv:2303.10130) · Anthropic Economic Index 2026 (CC-BY) |
how this is calculated |
last updated 2026-07-16