AI exposure: Paralegals and Legal Assistants
Assist lawyers by investigating facts, preparing legal documents, or researching legal precedent. Conduct research to support a legal proceeding, to formulate a defense, or to initiate legal action.
Reading this score
computedParalegals and Legal Assistants is one of the unusual cases where capability and friction are both high. Current systems can produce a great deal of this work, with average capability across its tasks at 2.5 out of 4. Only 33.6% of the weighted task load comes out as exposed, because 32.8% of it runs into something structural. Jobs in this band tend not to disappear. They change shape, and the person stays.
What holds the line here is accountability. Across this occupation's 12 tasks it averages 2.17 out of 3, the highest of the five friction dimensions. In plain terms, someone licensed has to sign the work. Capability and permission are different things. Where a statute or a professional body requires a named, liable human to approve the output, the work can be drafted by a machine and still require the person.
The most exposed thing this job does is Direct and coordinate law office activity, at 50.0%. The least is Call upon witnesses to testify at hearings, at 1.7%. A gap of 48.3% between two parts of the same job is the reason this index publishes at task level. An occupation-wide number would have hidden both.
Within legal occupations, this one is more exposed than most. The median across the 7 roles in the group is 20.1%, and only 2 of them score higher than this. Occupational families are not uniform, and the spread inside them is often wider than the gap between them.
What would move this score. Of 12 tasks, 5 are currently banded exposed, 4 assisted and 3 untouched. For that distribution to shift materially would take a change in who is permitted to sign the work, which is a question for regulators rather than for engineers. The score is re-computed every quarter against a fresh capability reference, and the change is published rather than quietly applied.
Where the score comes from
judgedEvery task is scored through the standardised work activities it maps to. These are this occupation’s averages on the six rubric dimensions. Capability is what AI can do; the other five are what stands in the way.
| Dimension | Mean | Scale |
|---|---|---|
| Capability | 2.47 | 0-4 |
| Embodiment | 0.86 | 0-3 |
| Presence | 0.94 | 0-3 |
| Accountability | 2.17 | 0-3 |
| Context | 1.96 | 0-3 |
| Verification cost | 2.17 | 0-3 |
What this means in practice
Roles in this band tend to change shape rather than disappear. The output gets drafted faster and the person moves toward review, judgment and accountability. The useful question is not whether to use these tools but who is trusted to sign off on what they produce.
Task by task
12 tasks, O*NET 31.0| Task | Exposed | Assisted | Untouched | Importance | Band |
|---|---|---|---|---|---|
| Direct and coordinate law office activity, including delivery of subpoenas. | 50.0% | 25.0% | 25.0% | 3.33 | exposed |
| Keep and monitor legal volumes to ensure that the law library is up-to-date. | 50.0% | 25.0% | 25.0% | 2.67 | exposed |
| Prepare affidavits or other documents, such as legal correspondence, and organize and maintain documents in paper or electronic filing system. | 42.5% | 32.5% | 25.0% | 4.38 | exposed |
| Prepare for trial by performing tasks such as organizing exhibits. | 40.0% | 35.0% | 25.0% | 3.97 | exposed |
| Gather and analyze research data, such as statutes, decisions, and legal articles, codes, and documents. | 40.0% | 35.0% | 25.0% | 3.89 | exposed |
| Prepare, edit, or review legal documents, including legislation, briefs, pleadings, appeals, wills, contracts, and real estate closing statements. | 35.0% | 40.0% | 25.0% | 4.32 | assisted |
| Investigate facts and law of cases and search pertinent sources, such as public records and internet sources, to determine causes of action and to prepare cases. | 35.0% | 40.0% | 25.0% | 4.06 | assisted |
| Meet with clients and other professionals to discuss details of cases. | 29.2% | 33.3% | 37.5% | 3.93 | assisted |
| Arbitrate disputes between parties and assist in the real estate closing process, such as by reviewing title searches. | 18.3% | 31.7% | 50.0% | 2.99 | assisted |
| Appraise and inventory real and personal property for estate planning. | 14.8% | 26.9% | 58.3% | 2.90 | untouched |
| File pleadings with court clerks. | 13.8% | 23.8% | 62.5% | 3.89 | untouched |
| Call upon witnesses to testify at hearings. | 1.7% | 23.3% | 75.0% | 2.99 | untouched |
Task text and importance ratings sourced from O*NET 31.0. Shares computed. The occupation score is the importance-weighted mean.
Occupations either side of this one
The four closest scores in the same occupational family, then the four closest anywhere in the index.
Read this carefully. Exposure is not displacement. A high score means current AI systems can produce this work, not that anyone will stop paying a person to do it. Adoption depends on economics, regulation and inertia that this index deliberately does not model. How the score is built.