Knowledge base
Legal AI, answered for Australian & New Zealand lawyers
Clear, practical answers to the questions practitioners — and AI answer engines — ask about legal AI: how reliable it is, how it handles confidential client data, what the conduct rules require, and how to adopt it without the risk.
Reliability & accuracy
- Is AI legal research accurate for Australian law? General-purpose AI chatbots are unreliable for Australian legal research — they can invent cases, statutes and citations that look real but do not exist. Purpose-built legal AI trained on a maintained Australian and New Zealand corpus is far more accurate, but accuracy still depends on two things: whether every answer carries a verifiable source citation, and whether the system is built to decline rather than guess when no authority exists. A lawyer must verify each result against the primary source.
- What are AI hallucinations in legal work, and how do you avoid them? An AI hallucination is a confident, fluent output that is simply wrong — an invented case, a misquoted section, or a citation to authority that does not exist. In legal work the danger is that hallucinations look exactly like good work. You avoid them by using AI that cites a verifiable source for every claim, is built to decline rather than guess, and by checking each citation against the primary source before you rely on it.
- Does AI cite its sources in legal research? Some legal AI does; most general chatbots do not. General tools usually produce an answer with no verifiable source, which is why their legal output cannot be trusted. Purpose-built legal AI should return a citation to the specific authority behind every answer — the case, section or rule — so you can click through and confirm it. A source on every result is the single most important feature for legal AI you intend to rely on.
- Is AI legal research reliable for New Zealand law? Only if the tool actually covers New Zealand law — and many do not. General chatbots and even some global legal AI treat NZ as an afterthought, with thin or no dedicated New Zealand coverage, which makes their NZ output unreliable. For trustworthy NZ legal research, use AI that maintains a current New Zealand corpus as a first-class jurisdiction, cites a source on every answer, and is built to refuse rather than guess. Then verify against the primary source.
Confidentiality & security
- Can lawyers use AI without breaching client confidentiality or privilege? Yes — provided the tool is built for it. The risk is not AI itself but where your data goes: consumer chatbots may use your inputs to improve their models and may store them offshore, which can put confidentiality at risk. Legal-grade AI keeps client content under a contractual confidentiality and no-training commitment and stores it securely. Privilege is generally preserved when you use a confidential, contracted service provider with proper access controls — but the lawyer remains responsible for vetting the vendor.
- Where is your data stored when you use legal AI? (Data sovereignty for Australian firms) It depends entirely on the vendor. Many global AI tools store or process data in the United States or Europe, which can expose it to foreign disclosure laws. For Australian and New Zealand firms, data sovereignty means knowing where client content is stored, where it is processed, and who can be compelled to hand it over. Look for Australian data storage, clear contractual terms, an Australian-owned or operated provider, and — where it matters most — an option to keep AI processing onshore.
- Is client data used to train AI models? It depends on the tool, and the difference matters enormously for lawyers. Many consumer AI tiers reserve the right to use your inputs to improve their models, which is a poor fit for confidential client material. Legal-grade AI should commit, in its contract, not to use your content to train any model. Always confirm that no-training commitment in writing before putting client matter into any AI tool.
- What security should a law firm look for in a legal AI tool? Look for independent security credentials such as ISO/IEC 27001 certification and a SOC 2 attestation, encryption in transit and at rest, clear data-residency terms, a contractual commitment not to train on your content, role-based access controls, and a signed data processing agreement. Marketing claims are not enough — ask to see the certificates and the contract terms in writing.
Ethics & professional rules
- Is it ethical for lawyers to use AI? Yes — using AI is ethical, and engaging with it is increasingly part of competent practice, provided the lawyer stays responsible for the work. Regulators in Australia and overseas treat AI as a tool: the existing duties of competence, confidentiality, supervision and candour to the court still apply unchanged. The lawyer — not the AI — is accountable for the accuracy of what is filed and advised. Used with verification and proper data safeguards, AI is consistent with the professional-conduct rules.
- Do AI tools comply with the Australian Solicitors' Conduct Rules? AI tools are not compliant or non-compliant in themselves — the Rules bind you, not the software. The Australian Solicitors' Conduct Rules require competence, confidentiality, honesty and candour to the court, and proper supervision, and those duties apply to how you use AI. Used with verification of output, safeguards for client data and supervision, AI use is consistent with the Rules. The obligations sit with the solicitor.
- How do Australian courts treat AI-assisted or AI-generated work? Cautiously, and increasingly with explicit rules. Australian courts have not banned AI, but several have issued practice notes requiring lawyers and self-represented litigants to disclose AI-assisted material, and confirming that the usual duties of accuracy, candour and verification still apply. The consistent message is that you may use AI, but you remain responsible for everything you file — and fabricated AI citations will not be excused.
Capabilities & use
- Can AI draft legal documents? What are the limits? Yes — AI can produce strong first drafts of contracts, advice, correspondence and submissions, especially when it works from your precedents and your firm's style. The limits are judgment and verification: AI does not exercise legal judgment, cannot be left to settle strategy, and every clause and citation it produces must be checked. Used as a fast first-drafter under a lawyer's supervision it saves real time; used unchecked it is a liability.
- Can AI review contracts and run due diligence? Yes — this is one of AI's strongest legal use cases. AI can read a contract in seconds, flag risk against your usual position, propose redlines, and work across a data room of hundreds of documents to surface obligations, gaps and inconsistencies. It does not replace the lawyer's judgment on what matters, and its findings must be verified — but for the volume work in review and diligence, it turns days into minutes.
- Can AI build a chronology from large document sets? Yes — building a source-linked chronology from a large set of documents is one of the highest-value things legal AI does. It can read hundreds or thousands of documents, extract dated events, and assemble a chronology with each entry pinpoint-cited back to its source document. The lawyer still checks the entries and decides what is relevant, but the days of manual timeline-building collapse to minutes.
- How do you write effective prompts for legal AI? Be specific about the task, the jurisdiction and the output you want, give the AI the source material to work from, and ask it to cite its sources. Good legal prompting looks less like a search query and more like instructing a capable junior: state the matter type and jurisdiction, attach the relevant documents, define the deliverable, and ask for authority on every point. Then verify what comes back.
Adoption & choosing
- Will AI replace lawyers and paralegals? No — but it is changing what they spend their time on. AI is very good at the high-volume, repetitive work (reading documents, first drafts, research, chronologies) and poor at judgment, strategy, advocacy and the client relationship. The realistic picture is augmentation, not replacement: lawyers who use AI handle more work with less grind, while the judgment — and the responsibility — stays human.
- How can small and sole-practitioner firms adopt AI safely? Start with one high-volume task, choose a tool built for legal work with proper data safeguards, and keep a human verifying every output. Small firms actually have an advantage: no procurement committee, fast decisions, and the most to gain from automating the grind. The safe path is to pick a low-risk, high-frequency use case, prove it on real matters, and expand from there — with verification built in from day one.
- What should you ask a legal AI vendor before buying? Ask about accuracy (does it cite a source on every answer, and is it built to refuse rather than guess?), jurisdiction (does it cover AU and NZ law, maintained and current?), data (where is client content stored and processed, and is it used to train models?), security (ISO 27001, SOC 2, a DPA?), and commercial terms (pricing, trial, lock-in). Get the data and security answers in writing — not in a sales deck.
- What should Australian and New Zealand firms look for when choosing legal AI? Look for genuine AU and NZ jurisdiction coverage (maintained and current, not a global tool with a thin local layer), a citation on every answer, client data stored in Australia with clear processing terms, integration with the practice-management systems you already run, and transparent pricing with a trial. The best fit for an ANZ firm is generally a tool built around ANZ law and data from the start — not one adapted to it.