AI for lawyers has settled into a specific set of tasks.
The ABA's ethics opinion on generative AI names legal research, contract review, due diligence, document review, regulatory compliance and drafting among the tasks these tools can assist with, and it treats technology-assisted review of e-discovery documents as a well-known use.
This page covers which tool categories do that work, the ethics duties that attach, and the AI skills worth evidencing on your resume.
Which AI tools do law firms use?
Ask which AI tools law firms use and the durable answer is a set of task categories rather than a product list.
Formal Opinion 512 — the American Bar Association's ethics opinion on generative AI, issued July 29, 2024 — describes tasks these tools may assist with, naming legal research, contract review, due diligence, document review, regulatory compliance, and drafting letters, contracts, briefs and other legal documents among them.
- Legal research — finding and checking the law
- Contract review — first-pass review of agreements
- Due diligence — working through document sets in transactions
- Document review — high-volume review in litigation
- Regulatory compliance — tracking the rules that apply to a client's business
- Drafting — letters, contracts, briefs and other documents
In e-discovery, the opinion names technology-assisted review as a well-known AI use: lawyers use it to categorize vast quantities of documents as responsive or non-responsive and to segregate privileged ones.
High-volume document review in litigation is that workflow.
If you arrive asking about a product by name — Harvey AI, say — the answer above is still the one this page gives: the work, not a product roundup.
The task categories are the part that travels between firms.
A lawyer who can run research or review with verification habits attached can learn a new firm's platform on the job.
Ethics rules for AI: competence, confidentiality, supervision (Opinion 512)
The rules come from the states; the analysis comes from the ABA.
Formal Opinion 512 was issued on July 29, 2024 by the ABA Standing Committee on Ethics and Professional Responsibility, and it does not write new AI rules — it applies the Model Rules lawyers already follow to generative AI tools: Rule 1.1 competence, 1.4 communication, 1.5 fees, 1.6 confidentiality, and 5.1 and 5.3 supervision.
In plain terms, the opinion ties generative AI use to the duties of competence, confidentiality, client communication, supervision, meritorious claims, candor to the tribunal and reasonable fees.
Competence is the anchor: lawyers need not become AI experts — what the duty requires is a reasonable understanding of the capabilities and limitations of the generative tools they use.
Practically, that means knowing what a tool does well, where its output fails, and when a result needs independent checking before it reaches a client or a court.
Confidentiality is the duty with day-to-day consequences.
Under the opinion, a client's informed consent is required before inputting information relating to the representation into a self-learning generative tool — one that learns from what you feed it — and boilerplate consent in an engagement letter is not enough.
Fees follow the same logic.
A lawyer may not bill a client for time spent learning a generative tool the lawyer will use regularly for clients.
And when billing hourly, only actual time spent can be billed — a fee for little or no work is not allowed.
Supervision reaches the whole firm.
Managerial lawyers must set clear policies on the firm's permissible use of generative AI, and supervisors must make sure lawyers and staff comply with their professional obligations when using the tools — and that they are trained.
Model rules, state versions
How AI changes associate tasks
The tasks shift first.
Where a generative tool produces the first pass — a research memo, a contract summary, a draft — the associate's work moves around it: framing the question for the tool, checking what comes back, and taking responsibility for what gets filed or sent.
The opinion's hallucination warning is the reason why: some tools produce plausible-sounding output with no basis in fact, and uncritical reliance can mislead clients and courts.
The fee rules shape the economics.
Learning time for a tool you will use regularly is unbillable, and hourly entries cover actual time spent — so when a tool cuts a task from an afternoon to an hour, the entry is an hour.
On hourly matters, efficiency shows up in the client's bill, not on the timesheet.
Firms respond with policy and training: managers set the rules for permissible use, and supervisors make sure lawyers and staff follow them and get trained.
For the profession-wide picture, the Bureau of Labor Statistics expects some routine legal work to be automated but does not expect that efficiency to reduce overall demand for lawyers.
AI skills to put on your resume
The useful resume skills are the ones the ethics rules already test, because they are specific and verifiable.
A hiring partner cannot check "familiar with AI"; they can check whether you can state what a tool does well, where it fails, and what you do about the second part.
- Working knowledge you can state precisely — what your tools do well and where they fail; the opinion's competence standard is a reasonable understanding of capabilities and limitations
- A verification habit — citations and facts checked before anything leaves your desk, because generative output can be plausible and baseless
- Confidentiality judgment — you know what may go into a self-learning tool, and when a client's informed consent comes first
- Task categories, named concretely — legal research, contract review, due diligence, document review, regulatory compliance, drafting
- Policy fluency — you have worked under a firm's AI use policy, and you know supervision duties run to staff as well as lawyers
If a posting names a platform you have used, say so; otherwise describe the task you used a tool for and the check you ran on its output.
The task list transfers between firms; the vendor logo does not.
These skills sit alongside the fundamentals of attorney careers, not in place of them.
Hallucinated citations and sanctions: the career risk
The opinion warns that some generative tools hallucinate — producing output that looks plausible but has no basis in fact — and that uncritical reliance can mislead clients and courts.
The duties that risk engages are on the opinion's list: meritorious claims and candor to the tribunal.
The cautionary case is Mata v. Avianca.
In the Southern District of New York (No. 22-cv-1461, June 22, 2023), the court sanctioned two lawyers and their firm $5,000, jointly and severally, for submitting non-existent judicial opinions with fake quotes and citations created by ChatGPT — and for continuing to stand by the fake opinions after judicial orders called their existence into question.
The career lesson is not that AI tools are dangerous; it is that verification is the lawyer's job no matter which tool produced the draft.
A sanction order is a public court document, and the Mata order names the two lawyers and their firm.
Checking every citation before anything is filed is the cheapest career insurance on this page.
Where to find attorney jobs
The tasks above appear across attorney work — litigation, transactions, compliance, and the review-heavy teams in between — so the honest entry point for AI lawyer jobs is the ordinary attorney market, plus document review and e-discovery teams where the tooling is the workflow.
On attorney jobs at LawFirmHires you can browse current openings at law firms, with the task categories above as a checklist to read each posting against.
Career information, not legal advice. The rules described here are the ABA's model rules as interpreted by Formal Opinion 512; states adopt and change them. Confirm the current rules with your state's disciplinary authority or your state bar's ethics counsel before relying on them at work.