Community legal centres and law centres: practical uses of AI
By the ChamberOne team 5 minute read
A community legal centre has a phone that rings all morning, a waiting list that grows and a handful of staff and volunteers. Every minute spent typing a note or drafting a standard letter is a minute not spent with the next person. AI tools can give some of that time back. They can also, handled carelessly, put the most sensitive information in the wrong place. A centre needs to be clear about both sides before it starts.
Key points
- Use AI to support, not replace, the person who decides who gets help.
- Always keep a quick route to a human for anyone who may be in danger.
- Have a lawyer check plain-language materials and translations before release.
- Use the least personal data possible, and avoid free consumer tools for client material.
- Start with one small pilot, with clear success measures and supervisor review.
The setting: tight budgets, high stakes
Community legal centres in Australia and law centres in the United Kingdom serve people who often cannot afford a lawyer and whose problems, such as housing, debt, family breakdown, immigration and employment, may be urgent. Funding is usually restricted, staffing thin and volunteers numerous.
That shapes how AI should be considered. The possible gains are real, because small time savings multiply across many clients. But the people served are often vulnerable, the information is sensitive and the centre may have funders and regulators who expect high standards. Choose cautiously and start small.
Triage and intake
Most centres begin with a queue of enquiries that need sorting. AI can help staff do it faster, but it should support the person making the decision, not replace them.
- Summarising a long enquiry message into the key facts and the likely area of law, for a staff member to confirm.
- Flagging words that suggest urgency, such as a deadline, an eviction date or a safety concern, so that a person looks at those first.
- Suggesting which questions to ask on the phone, based on the type of problem.
- Drafting a first reply that says what the centre can and cannot help with, for a person to approve.
Do not automate the decision to turn someone away
A tool may misread an urgent case as routine, and a person with limited English or a difficult story may not be understood. Any decision to decline, delay or refer should be made by a person who has read the enquiry. Where a tool is used to sort, keep a way for someone to override its classification and to check a sample of the lowest-priority cases each week.
Never let a tool's automated message be the only response to a person who may be in danger. Make sure every route into the centre offers a way to reach a human quickly and tells people what to do in an emergency.
- Have a person who has read the enquiry decide to decline, delay or refer
- Let staff override the tool's classification
- Check a sample of the lowest-priority cases each week
- Offer every route a quick way to reach a human
- Let an automated message be the only reply to someone who may be in danger
- Let a tool decide to turn someone away
Plain-language information
Many centres produce fact sheets, letters and guides. AI can help draft and simplify them, with human experts keeping control of the content. A lawyer writes or approves the substance, and a tool can help reduce reading difficulty, suggest headings and propose a short summary.
Test materials with real readers where possible. A leaflet that reads simply to a lawyer may still confuse someone for whom English is a second language. If a tool is used to translate, have the result checked by a competent speaker before it goes out, because errors in legal translation can mislead.
Internal tasks that save volunteer time
Inside the centre, there are many safe uses. Turning rough case notes into a tidy file note. Preparing a chronology from a client's documents. Producing a first draft of a standard letter to a landlord, a creditor or an agency. Summarising a long decision for a volunteer who is new to the area. Preparing training material for volunteers.
Each is a draft for a lawyer or supervised volunteer to check. Make that rule explicit. Volunteers, particularly students, may trust a polished output more than they should. Give them a short guide on what typical errors look like and require supervisor review of anything that leaves the building.
Data protection and confidentiality
A centre holds data about people in distress. Before using any tool with client information, ask where the data goes, who can access it, whether it is used to improve the tool and how it is deleted. Free consumer tools are rarely suitable for client material. Check the position under the data protection law where you work and the guidance of your regulator, funder or professional body.
Use the least data you can. Remove names, addresses and dates of birth from material before using a tool for drafting or summarising. Do not put in information about children, health or immigration status unless the tool has been properly vetted. Record in a short register which tools are approved and for what.
Quality controls on a small budget
Quality control need not be expensive. A few habits make the difference.
- Every AI-assisted document is read in full by a lawyer or supervised volunteer before it is sent.
- Every legal statement is checked against an official source, such as legislation or a regulator's website.
- A monthly spot-check of a few files looks at how tools were used and what errors were found.
- A simple log records errors caught, so that the centre learns from them.
- Staff and volunteers are told that they may raise concerns about a tool without blame.
Starting well: a modest pilot
Choose one task, such as file notes for a single service, and trial it for a month with a small group. Agree in advance what success looks like: time saved, errors found, staff views. Involve the people doing the work in the design, and tell clients, in plain words, if a tool is used on their matter. At the end, decide whether to continue, change or stop. Share what you learn with other centres, since many face the same questions and few have time to test alone.
Funders may ask how technology is being used. A short, honest account of what you do and what controls apply is better than silence, and it helps to keep trust with the people the centre serves.
This article is general information about the law, legal practice and technology, current when it was published. It is not legal advice. Laws and guidance change: check the sources linked above and take advice on your own circumstances from a lawyer qualified where you are.
Read next
AI and access to justice: an honest assessment
Where AI can lower the cost of routine legal help, where it cannot replace advice, and who bears the risk when it is wrong.
5 minute readClient portals: what self-service does for clients and the firm
Which questions a portal answers, what to keep out of it, and how it reduces routine calls.
5 minute readAI chatbots on law firm websites: benefits, risks and rules
A chatbot can help with intake, but it can also say too much. Decide its limits before it goes live.
4 minute read