Free sampler. Re-scored every six months.
Law scores between 4.5 and 7.9 out of 10 for AI exposure depending on the work, where 10 is most at risk. Trial advocacy scores 4.5. Junior associate and document-review work scores 7.9 — the widest internal split of any licensed profession we track. Legal employment is at a ten-year high while the route into the profession narrows. Both of those things are true at once.
The short answer for parents: law still leads somewhere valuable — but the old bargain has broken. Survive three expensive years, grind through document review, emerge a lawyer. That deal assumed the document-review years would exist to grind through, and increasingly they don't.
Every career in this index is scored 1–10, where 10 is most exposed to AI. Same six factors, same weights, applied identically to a litigator and a paramedic.
| Law track | 2023 | 2025 | Now | 3-yr move | Band |
|---|---|---|---|---|---|
| Trial litigator / advocacy | 3.8 | 4.1 | 4.5 | +0.7 | Low–Mod |
| Regulatory & compliance counsel | 4.1 | 4.5 | 5.0 | +0.9 | Moderate |
| In-house counsel | 4.3 | 4.8 | 5.2 | +0.9 | Moderate |
| Transactional / corporate associate | 5.1 | 5.7 | 6.3 | +1.2 | Mod–High |
| Junior associate / document review | 6.0 | 7.2 | 7.9 | +1.9 | High |
2023 and 2025 figures are reconstructed using current methodology, not archived from past editions.
For scalethe median career in this edition scores around 5.5. Entry-level software development scores 8.1. Bedside nursing scores 2.8.
The junior track is the fastest-moving score in the entire index — up 1.9 points in three years, faster than entry-level software development.
The split is 3.4 points wide. Same bar card, same law school, same firm. What differs is whether the work is producing documents or standing behind a decision.
Not the ones who go to court, advise on risk, or sign their name to advice. It has already replaced a great deal of what junior lawyers were hired to do.
The pattern in law is unusually stark because the automatable layer and the training layer were the same layer.
How a junior associate rates against each. Ratings are 0–10 on each factor's own terms.
So you can see what the analysis actually looks like.
Does the law require a licensed human? — rated 7.0
Does the law require a licensed human? — and why the bar card doesn't save everyone
Licensing is normally the most durable protection in this index. It moves at the speed of legislation rather than technology, and no advance in capability repeals a scope-of-practice rule.
Law has it in full. Only a barred attorney can appear, advise and sign. That is why trial advocacy rates 9.5 on this factor and sits comfortably in the low band.
But protection attaches to acts, not to people. The bar card protects the appearance in court, the signed advice, the fiduciary relationship. It does not protect the reading of forty thousand documents. Nothing in any statute says a human must perform first-pass discovery review — so nothing prevented that work from being automated the moment the technology was adequate, which it now is.
This is why a junior associate rates only 7.0 here rather than 9.5. They hold the same license as the partner. The license simply doesn't reach the work they are doing.
The lesson generalizes beyond law: ask what a licence actually reserves, not whether one exists. Nursing's licence reserves clinical practice and mandates the training hours, which is why nursing's on-ramp is protected. Law's licence reserves the courtroom and the signature, and leaves the apprenticeship exposed.
The honest answer is that it depends on a number most law schools do not publish clearly.
Legal employment in the United States hit a ten-year high in early 2026. The profession is healthy. But only 38% of 2025 associate hires came directly from graduating classes, down from 46% two years earlier, while lateral hiring of experienced lawyers rose. Firms are buying judgment rather than building it.
At the same time, paralegal hiring at the largest firms has fallen sharply since 2018, and a large share of paralegal tasks are automatable with current technology.
So the profession is expanding while its base is narrowing. For someone already inside, that is a good market. For someone trying to get in, it is a harder one than the headline employment figures suggest — and law school debt is priced on the old assumption. Georgetown's analysis of 186 law schools found median earnings net of debt of $72,000 four years out, but under $55,000 at 33 of them.
Ranked by exposure, safest first:
The consistent shape: exposure falls as accountability rises. The further a role sits from someone personally answering for the outcome, the more exposed it is.
We think it deserves stating plainly.
Cutting junior hiring may be a self-defeating mistake rather than a permanent shift. Every senior lawyer was once a junior, and there is no alternative factory producing experienced ones. On this view firms will be forced to rebuild the pipeline when the shortage arrives, and today's squeeze is a painful cycle rather than a new normal.
Some firms are already reserving work for manual completion specifically to preserve its training value — which is a profession admitting out loud that it has a pipeline problem.
We think the displacement is structural, and we score the market as it is rather than as it ought to be. But if graduate hiring rebounds, our score should come down and we will say so. The full profile sets out both cases with sources.
This sampler tells you where law stands. The full profile tells you what to do about it.
Most families are weighing two or three careers seriously, and a few more they haven’t ruled out. Pick the ones you need.
Each includes a short version written directly to the student and the technical scoring appendix. Spring 2027 updates of whatever you buy are included.
Read one complete profile free → We publish computer science in full so you can judge the depth before buying anything.