The fastest way to get AI wrong as an Australian lawyer isn’t a bad model — it’s forgetting that your professional duties don’t change just because a machine drafted something. The good news: the obligations are clear, and meeting them is entirely compatible with using AI well.
What the guidance converges on
State law societies and regulators have issued increasingly specific guidance on AI use by solicitors. Across NSW, Victoria and Western Australia, the common threads are:
- Confidentiality. You remain responsible for client information. That rules out feeding matters into tools that train on your inputs or process them offshore by default.
- Competence and supervision. AI output is a draft to be checked by a lawyer, not an answer to be passed on. You’re accountable for what leaves the firm.
- Candour and accuracy. Hallucinated citations have already embarrassed lawyers overseas. Verification isn’t optional.
- Disclosure where appropriate. Know when clients or courts should be told AI was involved.
Translating duties into setup
These obligations point to a specific kind of deployment:
- Tools that run with Australian data residency and don’t train on your data — the heart of our piece on data sovereignty.
- Output with sources you can verify, so candour and competence are achievable in practice.
- A human in the loop by design, not as an afterthought.
- Tooling tuned to Australian law, because accuracy in a local jurisdiction is where generic AI falls down.
The point
Your duties aren’t a reason to avoid AI. They’re a specification for how to adopt it. Firms that build to that spec get the productivity and keep the trust their clients rely on.
If you want help adopting AI in a way that fits your obligations, book a call.