The core unit of analysis is the task inside a real workflow, not a prediction about whether an occupation will “be automated.”
Unit of analysis
Roles are treated as changing bundles of tasks, decisions, relationships, responsibilities, knowledge, and context. A workflow is examined for purpose, inputs, outputs, judgment, exceptions, consequence, evidence, reversibility, and human accountability.
Each task is placed in one of three provisional lanes: human-led, AI-assisted, or delegated with review. A lane is not permanent. It can change as evidence, capability, policy, or consequence changes.
Evidence labels
Observed evidence
A finding reported by a linked research source. We preserve the population, timing, and limits needed to interpret the finding.
Operating guidance
A practical recommendation drawn from an established framework, standard, or regulator guidance. It still requires local adaptation.
Design hypothesis
A task-level starting point offered for discussion and testing. It is not a factual prediction, policy determination, or employment recommendation.
Role playbooks
Playbooks use occupation and task foundations from O*NET, then add work-design hypotheses informed by current AI capability patterns, the NIST AI Risk Management Framework, workforce research, and practical review questions. They are intentionally concise so a team can challenge and adapt them.
Limits and updates
AI capability, law, standards, products, and organizational context change. Research pages display a review date. Material claims link to the source used. When a source changes or stronger evidence appears, guidance should be re-evaluated.
This site is an educational resource. It does not provide legal, employment, labor, safety, medical, privacy, cybersecurity, financial, or regulatory advice. Qualified review is necessary for consequential use.
To raise a correction or source concern, use the contact channel on Mojica Consulting and identify Human AI Workforce and the page in question.