OBSERVED IN REAL USE · Anthropic 2026
19%
of this role’s work is already showing up in real Claude usage (Anthropic Economic Index).
You might want to sit down for this one.
77% 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. Deeply cooked. Almost the entire task list is something a model can already attempt.
WHAT AI CAN ALREADY DO
- Resolve discrepancies in accounting records
- Prepare itemized statements, bills, or invoices and record amounts due for items purchased or services rendered
- Post stop-payment notices to prevent payment of protested checks
- Keep records of invoices and support documents
- Perform bookkeeping work, including posting data or keeping other records concerning costs of goods or services or the shipment of goods
WHAT IT STILL CAN’T
- Monitor equipment to ensure proper operation
- Fix minor problems, such as equipment jams, and notify repair personnel of major equipment problems
- Weigh envelopes containing statements to determine correct postage and affix postage, using stamps or metering equipment
- Load machines with statements, cancelled checks, or envelopes to prepare statements for distribution to customers or stuff envelopes by hand
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 77% 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