Menu
Lawyer
Three components — Automation Resistance, Structural Moat, and Demand — add up to the 60.
Licensed responsibility is the durable part: advice, filings, advocacy, negotiation, and client risk still need a lawyer who owns the answer. Research, document review, cite checks, and drafting are the exposed layer, especially in first-year training.
Observed AI overlap is meaningful, and modeled job-loss risk puts real pressure on junior research, drafting, cite checking, contract review, and due diligence. The work holds better only when a lawyer is carrying client judgment, courtroom or negotiation responsibility, and a licensed signature that has consequences.
Research/drafting AI; partner/solo billing captures, associate hours compress. The tools raise output, but the worker-side payoff depends on ownership, billing power, book of business, or senior responsibility; otherwise the employer or platform captures much of the gain.
The state bar is the dominant protection: law degree, bar exam, ethics rules, continuing duties, and unauthorized-practice limits create a stronger gate than most knowledge jobs carry, even though independent authority builds with practice and trust.
Office/court/client; lift 7.21 lb, standing 13.4%, outdoor 8.9%. The work is mostly office, client, court, or field-inspection work rather than physically demanding labor, so physical conditions add little protection against software substitution.
State bar admission (J.D. + bar exam + UPL enforcement) gates legal practice with independent scope. This matters only where the rule actually gates practice; voluntary credentials and market signals help, but they do not protect the occupation the way a required license does.
No robotic path for legal work. The substitution story is software and workflow automation, not machines taking over the job site, office, court, or client conversation.
Job Zone 5; doctoral/professional (J.D.) entry. Training time creates some screening power, especially when the path includes a degree, exam, supervised work, or respected professional credential.
Legal hiring is steady and resilient where disputes, compliance, transactions, and risk advice require licensed judgment, but growth is modest and some document-heavy work can be compressed by AI. The bar still separates practice from support.
SOC 23-1011: 4.1% growth, 31.5k openings on 864.8k. The volume score reflects both the size of the workforce and the number of annual openings, not just whether the occupation is growing.
Professional-services demand + replacement. Demand is stronger when it comes from durable business, legal, financial, or client need; it is weaker when it depends mostly on churn, cost-cutting, or work that software can absorb.
Bar monopoly blocks role substitution; AI compresses research/drafting core and entry cohort. The key question is whether the human part remains necessary as AI tools improve; lawyer keeps some protected work, but the early or routine layer still needs watching.
If judges and state bars keep sanctioning unchecked AI citations and require clearer lawyer review, the licensed-responsibility moat strengthens. That would protect lawyer work even while tools keep speeding research and drafting, especially in litigation, regulated advice, and signed filings.
If employers rely on legal-AI systems to cover more first-pass research and document work, the first-job funnel narrows even if senior legal demand holds. That would hurt the path most for new graduates carrying high debt and weak placement odds.
If states let nonlawyers or AI-first services handle more routine legal matters without close attorney responsibility, the bar moat weakens at the consumer and small-business end. Complex litigation, regulated counsel, signed court work, and high-stakes negotiations would be less affected.