- PDF doc
- 297.92 KB
This checklist is designed to help you comply with your professional obligations and avoid common problems associated with AI (artificial intelligence) use.
The checklist is provided in text below, as well as downloadable Word and PDF versions. If you would like to adjust the Word version to suit your needs (e.g. to add additional notes), open the document in Word and go to Developer > Restrict Editing > Stop Protection to unlock the document for editing.
Use this checklist in combination with our other AI resources:
- AI for lawyers: ways to use AI in legal practice
- Statement on the use of artificial intelligence in Australian legal practice
- AI checklist for law practice leaders and principals
For more information and resources visit our AI web page.
Understand your professional and ethical obligations
The Legal Profession Uniform Law Australian Solicitors’ Conduct Rules 2015, Legal Profession Uniform Conduct (Barristers) Rules 2015, and obligations under the Legal Profession Uniform Law apply to AI use. Make sure your use of AI complies with these obligations.
Understand your organisation’s AI governance and policies
If you’re a law practice leader and need to develop your practice’s AI governance, see our AI checklist for law practice leaders and principals for guidance.
Confirm whether, and how, you are authorised to use AI
Check what:
- tasks you are permitted to use AI for
- tools you are permitted to use for those tasks
- information is approved for use with those tools.
Never use client information or other confidential, personal or private information (whether in prompts or uploaded documents) with an AI tool that hasn’t been approved by your practice for that use.
Entering confidential, personal or private information into an insecure AI tool can breach your professional obligation to maintain client confidentiality, as well as other privacy and confidentiality obligations.
Confirm AI can be used on the client’s matter
Confirm that the use of AI is consistent with your client’s instructions and the terms of the retainer.
Consider whether the client should be warned about their own AI use
Clients may not realise that entering confidential or privileged documents into AI tools can create risks for confidentiality, privilege and privacy, and that AI-generated material may be unreliable or inaccurate.
Verify all AI-generated materials
Do this before relying on or sending AI-generated materials. See section 3 for guidance on how to verify.
Keep accurate records of how AI has been used on a matter
Include:
- time spent
- what the AI was used for
- how any AI-generated material was verified.
If something goes wrong, someone (a court, another party, a client, your supervisors or a regulator) may want answers. Your practice should be able to explain whether you used AI, how it was used, what steps were taken to verify the work, and what sources you relied on.
Check for any other information you are required to record by your workplace policy and consider saving copies of any GenAI prompts and responses.
Make sure your supervisors know how you are using AI
They’ll tell you which AI tools and information you can use and direct you to relevant workplace policies and training.
Before relying on AI-generated materials or sending them to others
Check:
- the primary sources, such as legislation, cases or evidence – do not rely only on AI-generated citations, summaries or source links
- all quotations, references and citations are accurate, exist, and support the proposition for which they are cited
- that the output applies to the correct jurisdiction – AI may draw on material from other jurisdictions
- that the output is current – AI may be trained on out-of-date information
- that the output is complete – AI may omit important information or may not have access to relevant information held by your practice
- for bias or unsupported assumptions, and other errors including hallucinations.
GenAI (and other AI tools based on machine learning) can produce convincing but false or misleading information (often called hallucinations). Examples include non-existent cases, parties or precedents, legal analysis that does not make sense, incomplete answers, or citations that do not support the proposition for which they are used. This is because it works by identifying patterns and making statistical predictions, not by understanding meaning or checking what is true.
Effective prompting and technical features (like ‘retrieval augmented generation’) may improve outputs, particularly where a tool draws on legal databases or your practice’s own materials, but they do not remove the risk of error or hallucination.
Always verify AI outputs before relying on them. Failing to do so may breach your professional and ethical obligations, harm your client’s case, and may lead to judicial reprimand or disciplinary action.
Consider using a structured self-review process, such as reviewing the material again after a break, and recording the verification steps taken.
When receiving materials from others
Check:
- whether AI was used to prepare them
- whether and how they were verified.
Checking how materials prepared by others were developed reduces the risk that you inadvertently rely on unverified AI-generated material.
Bill clients appropriately
Do not bill for:
- unnecessary costs caused by using AI
- time spent learning to use AI.
Costs must be fair, reasonable and proportionate. Know the reasonable costs of doing a task without AI. AI should not unreasonably increase the costs for clients, for example because of extra time spent verifying or correcting AI outputs.
Time billed to your clients must also accurately reflect the time spent on their matter, not your professional development.
Disclose to clients how you have used AI on a client’s matter
Follow relevant:
- instructions from your client
- disclosure steps required by your workplace’s AI policy.
Make sure materials have been verified before they are submitted to a court or tribunal
Submitting AI-generated material to a court or tribunal without proper verification can seriously harm your client’s case. It may include false citations or incorrect statements of fact or law. It may also waste court resources, damage public confidence in the justice system, and lead to disciplinary action.
For litigation documents, consider making a file note at the time recording which parts of the document were AI-assisted, who verified them and how they were checked. Some courts or tribunals may require you to identify or explain AI use, and a contemporaneous note will help you respond if that information is requested.
Identify any court or tribunal requirements
Different jurisdictions have different requirements for AI use in litigation. Check for any relevant:
- court rules
- practice notes
- guidelines.
Check requirements before using AI with witness information
Follow any applicable court or tribunal rules, practice notes or guidance.
Law Library Victoria’s AI in the Law for Legal Practitioners guide has a list of Australian court and tribunal guidelines, and practice notes you can refer to. Make sure to check for any updated versions from the relevant court or tribunal.
Keep in mind that relevant requirements may be in documents that aren’t specific to AI (e.g. rules about expert witness reports or affidavits).
Comply with any disclosure requirements
Some practice notes or rules require disclosing certain uses of AI.
Consider whether voluntary disclosure is appropriate
For example, to avoid misleading the court or another party.
Bring defective or non-compliant documents to the court’s attention
If you are responsible, take accountability and give a full and honest explanation at the first opportunity.
- Word doc
- 3.23 MB