AI can automate or assist parts of a lawyer's work, including drafting, document review and research. Whether that leads to fewer legal jobs, different roles or more legal services is uncertain. The effects will vary by practice area, client demand, pricing and how firms adopt the tools. Lawyers remain responsible for their advice and work product, but that responsibility does not guarantee that staffing, pay or career paths will stay the same.
What the research says
In March 2023, a Goldman Sachs report estimated potential AI exposure equivalent to 44% of work in U.S. legal occupations. This was an estimate based on occupational tasks and assumptions about AI capabilities, not a measurement of jobs lost or a prediction that 44% of lawyers would be replaced. Exposure alone does not tell us whether a task can be automated reliably, how quickly a firm will adopt AI or what happens to employment.
The International Labour Organization's 2025 occupational-exposure study concludes that job transformation is the most likely overall impact because most occupations still contain tasks requiring human input. That is a broader labor-market assessment, not a forecast of legal hiring or a promise that every existing role will remain.
Thomson Reuters' Future of Professionals 2025 report projects that AI could free up nearly 240 hours a year for legal professionals, roughly five hours a week. These are projections informed by a survey, not measured productivity gains or evidence about future headcount. Saved time could support more client work, lower prices or reduced staffing; the outcome depends on how organizations and clients respond.
The American Bar Association's Formal Opinion 512 applies existing ethical duties to lawyers' use of generative AI, including competence, confidentiality and supervision. These duties explain who must take responsibility for the work. They do not settle the question of how many lawyers a firm will employ. Check the rules and guidance that apply in your jurisdiction.
Which legal tasks can AI help with?
Common uses include the tasks below. Performance depends on the model, source material and workflow, so test each use on representative matters and include the time needed to check and correct the result.
- First drafts from your own precedent and templates.
- Summaries of long documents, transcripts and email threads.
- Reviewing contracts against your standards and flagging deviations.
- Finding the clause, the precedent or the matter detail across thousands of files.
- Extracting dates, obligations and parties into a table.
- Answering routine questions from your own policies.
- Filling repetitive forms and questionnaires from known facts.
What still needs a lawyer
When AI supports a lawyer's work, an accountable lawyer still needs to exercise and supervise the following responsibilities. How much time they require may change with the tools and the matter.
- Judgment: what the client actually wants, how much risk they will accept, and which argument is worth making.
- Responsibility: signing filings, giving advice and answering for it to the client, the court and the bar.
- Verification: AI output can be confidently wrong. Checking it is a lawyer's job. See how to avoid AI hallucinations in legal work.
- Negotiation and relationships: reading the other side, building trust with clients, managing a deal or a dispute.
- Unfamiliar questions: testing reasoning against the facts and applicable law, recognizing uncertainty, and deciding when to seek specialist advice.
- Ethics: conflicts, confidentiality and privilege decisions.
How legal work is changing
A workflow may start with an AI draft for a lawyer to review. That can move time toward checking sources, resolving exceptions and advising clients, provided the draft is useful enough to justify the review. It can also remove some of the routine research and drafting through which junior lawyers learn. Firms need to design training alongside automation.
Lower costs could let teams serve more clients or handle previously unmet demand. They could also reduce demand for particular tasks, affect hiring or change what clients will pay. Work designing, evaluating and supervising AI workflows may grow, but it is uncertain whether those roles will offset reductions elsewhere. Treat staffing and pricing decisions as questions to test against observed results.
Practical guidance for junior lawyers
- Keep learning the underlying work. Agree with a supervisor on opportunities to research an issue or draft a clause yourself, then compare your reasoning with the AI output and the final work product.
- Verify before relying on an answer. Open the cited source, check jurisdiction and currency, and test whether the passage supports the conclusion. Look for omitted exceptions and facts. Use our worked verification example.
- Use approved tools under supervision. Learn the tool's limits and your firm's rules for client data, permissions and review. Do not assume that a tool you use personally is approved for matter work.
- Keep a correction log. Record missed issues, unsupported claims and edits, and discuss recurring patterns with a senior lawyer. Measure whether the whole task becomes better or faster after review.
- Build judgment and client skills. Ask to join client discussions and negotiations, explain your reasoning, and seek feedback on decisions as well as drafts. Tool proficiency helps, but it is not a guarantee of job security.
Practical guidance for firm leaders
- Pilot a defined workflow. Establish a baseline and measure issue detection, source accuracy, corrections and total lawyer time. Set acceptance criteria before the pilot.
- Make review and accountability explicit. Name the responsible lawyer, define when approval is required, and specify which data and tools may be used. Adapt our free Word AI policy template and approved-tool worksheet to your jurisdiction and practice.
- Protect training opportunities. Preserve supervised drafting and research practice, teach source verification, and budget senior review time. Check whether juniors can explain the result without relying on the tool.
- Use evidence for staffing and pricing. Distinguish time saved on a task from capacity available across a team. Review client demand, fees, quality and workload before changing hiring targets or assuming that faster generation supports fewer staff.
- Plan for affected roles. Discuss changes openly, offer training and redeployment where feasible, and monitor workload and career progression. Reassess the workflow when models, documents or client needs change.
Where Gist fits
Gist supports first drafts from your own precedents, reviews against your standards and answers sourced from matter files. Its deployment team of lawyers and AI engineers builds and maintains agents for workflows specific to your team. Choose a defined task, test the results and agree on lawyer review before expanding use. See how Gist works or book a demo.
More from Gist
- Free AI Policy Template for Law Firms
- How to Avoid AI Hallucinations in Legal Work
- Generative AI vs Agentic AI for Legal Work
- Which AI Tools Review Contracts Inside Microsoft Word?
- Can Legal AI Draft Using My Firm's Own Precedents and Templates?
- Who Builds and Maintains Custom AI Agents for Law Firms?
- Gist Legal AI Evaluation: Contract Processing and Legal Research
