AI tools can save Australian lawyers significant time and help them work more effectively. But they also introduce new risks — and some of those risks can have serious professional consequences.

Here are the ten most common AI mistakes Australian lawyers make, and how to avoid them.

1. Relying on AI-Generated Case Citations Without Verifying Them

AI tools can hallucinate case citations — producing references to cases that do not exist, or misattributing holdings to real cases. This is one of the most dangerous mistakes a lawyer can make with AI.

In several high-profile cases in the United States and Australia, lawyers have filed submissions citing AI-generated cases that did not exist. The consequences have included costs orders, professional conduct referrals, and significant reputational damage.

How to avoid it: Every case citation generated by an AI tool must be verified against a primary legal database (Westlaw Edge, LexisNexis, Jade, or AustLII) before being relied upon or included in any document.

2. Using AI Outputs Without Review

AI tools can produce plausible-sounding but incorrect legal analysis, misstate legislative provisions, and miss important issues. Using AI output without careful review is not consistent with competent legal practice.

How to avoid it: Treat AI outputs as a first draft, not a finished product. Every AI-generated document, research summary, or legal analysis must be reviewed and verified by a qualified lawyer before being used.

3. Entering Confidential Client Information Into General-Purpose AI Tools

Many general-purpose AI tools, including free versions of ChatGPT, may use inputs to train their models. Entering confidential client information into these tools may breach your confidentiality obligations under the Australian Solicitors' Conduct Rules.

How to avoid it: Review the data handling policies of any AI tool before entering client information. Use purpose-built legal AI tools with appropriate enterprise data handling agreements, or anonymise information before entering it into general-purpose tools.

4. Misunderstanding What AI Tools Can and Cannot Do

AI tools are not legal research databases. They do not search a database of cases and legislation — they generate text based on patterns in their training data. This means they can produce confident-sounding but incorrect statements of law.

How to avoid it: Use AI tools for what they are good at — drafting, summarisation, and generating starting points — and use proper legal research databases for primary research and citation verification.

5. Failing to Disclose AI Use When Required

Some courts and tribunals are developing rules that require disclosure of AI use in submissions and documents. Failing to comply with these rules can have serious consequences.

How to avoid it: Monitor the rules and practice directions of the courts and tribunals in your jurisdiction. Where disclosure is required, comply with it.

6. Over-Relying on AI for Complex Legal Analysis

AI tools perform well on routine tasks but can struggle with complex, nuanced legal analysis — particularly in areas where the law is unsettled, jurisdiction-specific, or requires the exercise of professional judgement.

How to avoid it: Use AI tools to assist with routine tasks and as a starting point for more complex work. Do not rely on AI for the substantive legal analysis in complex matters without careful verification.

7. Not Supervising Junior Lawyers' Use of AI

If junior lawyers in your firm are using AI tools, the supervising lawyer remains responsible for the quality of the work. Failing to supervise AI use by junior lawyers is a supervision failure, not just an AI failure.

How to avoid it: Establish clear protocols for AI use in your firm, including what review process applies to AI-generated work. Ensure junior lawyers understand the limitations of AI tools and the verification required.

8. Using AI to Generate Witness Statements or Affidavits

Witness statements and affidavits must reflect the witness's own words and recollection. Using AI to generate or substantially rewrite a witness statement raises serious concerns about authenticity and may breach court rules.

How to avoid it: AI tools can assist with formatting and structuring witness statements, but the content must come from the witness. Do not use AI to generate the substance of a witness statement.

9. Ignoring the Billing Implications of AI

If AI tools significantly reduce the time taken to complete a task, billing the same amount as if the task had been done manually may not reflect the value delivered to the client and could raise concerns about overcharging.

How to avoid it: Consider whether your billing practices need to be updated to reflect the use of AI tools. Transparency with clients about the use of AI and the basis for your fees is good practice.

10. Failing to Stay Current With AI Developments and Guidance

AI technology and the regulatory environment around its use in legal practice are both developing rapidly. Lawyers who do not stay current risk falling behind on both the tools available to them and the obligations that apply.

How to avoid it: Monitor guidance from the Law Council of Australia, your state or territory law society, and the courts in your jurisdiction. Review your AI practices periodically as the technology and regulatory environment evolve.