Tech & AI, Explained

ai-for-lawyers

8 min read

AI for Lawyers — A Unified Australian Legal Practice Framework

Status: 🟩 COMPLETE 🟦 LIVING Section: decision-frameworks Tags: lawyers, legal-practice, decision, law-society, australian-law

The short answer

For Australian lawyers and legal professionals, AI is rapidly changing practice:

Strongly recommended:

  • Harvey or Spellbook for contract work
  • General AI (Claude, ChatGPT) for drafting routine communications
  • AI document analysis for case preparation
  • Legal research AI plus authoritative sources

Use carefully:

  • AI for legal advice (verification critical)
  • AI for case strategy (judgment matters)
  • AI processing privileged information

Critical Australian context: Legal Profession Uniform Law, state Law Societies, professional indemnity, confidentiality, hallucinated citations (Mata v. Avianca lesson).

Where AI genuinely helps lawyers

Contract work (the major use case)

  • Harvey AI — adopted by major Australian firms (Clayton Utz, Allens, Herbert Smith Freehills)
  • Spellbook — accessible mid-market option (works inside Word)
  • CoCounsel (Thomson Reuters)
  • Lexis+ AI

For:

  • Contract drafting
  • Contract review
  • Clause comparison
  • Risk identification
  • Negotiation preparation
  • Due diligence at scale

Document review and discovery

  • Large document collection analysis
  • Privilege review assistance
  • Pattern identification
  • Specific clause extraction
  • Translation of foreign-language documents

For investigations and large litigation: AI substantially accelerates discovery.

AI helps with:

  • Initial research direction
  • Concept explanation
  • Comparative analysis frameworks

Always verify with authoritative sources:

  • AustLII (austlii.edu.au) — Australian case law and legislation
  • JADE (jade.io) — case law and commentary
  • LexisNexis Australia
  • Westlaw AU
  • Thomson Reuters
  • CCH (Wolters Kluwer)

Critical: AI hallucinates case citations. Every cite must be verified against actual case databases.

Client communications

  • Initial response drafts
  • Status updates
  • Outcome explanations
  • Bills and fee discussions
  • Difficult conversations
  • Marketing for the practice

Internal practice work

  • Standard operating procedures
  • Compliance documentation
  • HR materials (with HR considerations)
  • Practice management
  • Training materials

Drafting routine documents

  • Standard agreements (with review)
  • Letters of demand
  • Caveats
  • Notices
  • Court documents (with strict review)

Trial preparation

  • Document organisation
  • Theme development
  • Cross-examination preparation
  • Argument structuring
  • (Strategy remains lawyer’s judgment)

Submissions and pleadings

  • Structure suggestions
  • Authority compilation (verify all)
  • Drafting assistance
  • Editing support

The critical risk: hallucinated citations

This is the most important risk:

The Mata v. Avianca lesson (2023)

US case where lawyer submitted brief with AI-generated case citations that didn’t exist. Result: sanctions, professional discipline, career damage.

Similar Australian cases

Various Australian incidents have emerged of lawyers and self-represented litigants submitting AI-generated content with fake citations.

Why it happens

  • AI generates plausible-sounding case names
  • Plausible-sounding judges
  • Plausible-sounding holdings
  • All fabricated

How to prevent

Verify every citation:

  1. Copy case name from AI
  2. Search AustLII
  3. Search JADE
  4. Verify case exists
  5. Verify it says what you claim
  6. Read the actual decision

If you can’t verify: Don’t cite.

This is non-negotiable. Career, client interests, and court integrity at stake.

  • NSW, VIC (and Western Australia 2022+)
  • Specific professional standards
  • AI use must comply

Other states

  • QLD, SA, WA, TAS, NT have own legal profession legislation
  • Different specifics
  • Similar broad principles

Law Society guidance

Each state Law Society has AI guidance:

  • Law Society of NSW
  • Law Institute of Victoria
  • Queensland Law Society
  • Law Society of Western Australia
  • Law Society of South Australia
  • Law Society of Tasmania
  • NT Law Society
  • Law Society of the ACT

All have published AI guidance (2023-2026). Familiarise with yours.

Law Council of Australia

Federal-level guidance on AI in legal practice.

Australian Bar Association

For barristers, additional considerations.

Federal Court and state courts

Various courts have published AI-related directions:

  • Disclosure requirements
  • Citation verification expectations
  • Specific practice notes

Specific issues

  • Privilege — legal professional privilege considerations
  • Confidentiality — client information protection
  • Conflicts of interest — AI doesn’t know your conflicts
  • Competence — must maintain professional skills
  • Supervision — junior lawyer + AI requires supervision

Professional indemnity

Lawyers’ insurance and AI:

Solicitors’ insurers

Lawcover (NSW), LawAssist, LPLC (Victoria), etc.

Considerations

  • Notify your insurer about AI adoption practices
  • Some claims may have AI-specific considerations
  • Document AI use practices
  • Stay within demonstrated competence

Particular risks

  • Hallucinated citations leading to negligence claims
  • AI errors in advice
  • Confidentiality breaches
  • Scope creep into areas outside competence

Solo and small firm

  • Spellbook for contract work (accessible)
  • General AI for client communications
  • Karbon Law or similar for practice management
  • Focus on highest-value work

Mid-tier firms

  • Harvey at some firms; Spellbook commonly
  • Document review AI for litigation
  • General AI for everyday
  • Practice management AI

Major firms (Big Law)

  • Harvey widely adopted
  • Custom AI deployments
  • Investment in AI infrastructure
  • Specialist AI teams

In-house counsel

  • Practical AI for daily work
  • AI for contract negotiation support
  • Communications AI
  • Specific corporate counsel AI emerging

Barristers

  • AI for research support
  • Document analysis
  • Argument structuring
  • (Strategy and advocacy remain personal)

Specific practice areas

Commercial law:

  • Strong fit with contract AI
  • Due diligence AI
  • Document analysis

Litigation:

  • Document review AI
  • Submission drafting
  • Citation verification critical

Family law:

  • Particularly sensitive client information
  • Cultural and emotional context
  • Communication assistance
  • (Strategic decisions remain personal)

Criminal law:

  • Case preparation
  • Evidence review
  • Defendant communications
  • Sensitive — privacy considerations crucial

Immigration law:

  • Multi-language client communication
  • Complex regulations
  • Document preparation
  • Strong fit for AI assistance

Personal injury / employment law:

  • Document analysis
  • Standard drafting
  • Client communications

Property and conveyancing:

  • Routine drafting
  • Compliance documentation
  • (Specific transactions still need lawyer attention)

Wills, estates, succession:

  • Drafting assistance
  • Communication
  • (Personal matters need careful approach)

Government and public sector law:

  • Policy analysis
  • Submission drafting
  • Briefing notes
  • (Sensitive information requires care)

Real workflows

Contract review with Spellbook

  1. Receive client contract for review
  2. Open in Word with Spellbook
  3. AI flags unusual or concerning clauses
  4. AI suggests negotiating positions
  5. You exercise legal judgment on what matters
  6. Communicate to client
  7. Negotiate

Litigation discovery

  1. Receive large document collection
  2. AI assists with categorisation
  3. AI flags potentially privileged documents
  4. Lawyer verifies privilege claims
  5. AI helps theme identification
  6. Lawyer makes strategic decisions
  1. Identify research question
  2. AI helps frame approach
  3. Search AustLII, JADE, Lexis, Westlaw for authoritative sources
  4. Read actual cases
  5. Verify every citation
  6. Apply legal reasoning
  7. Document research

Client letter drafting

  1. Notes from consultation
  2. AI generates draft based on facts and law
  3. Lawyer reviews and verifies legal positions
  4. Personalise for client
  5. Send

Court document drafting

  1. Outline key points
  2. AI assists with structure
  3. Lawyer drafts substance
  4. Lawyer verifies all citations
  5. Lawyer takes professional responsibility
  6. File

Confidentiality and AI

Lawyers have heightened confidentiality obligations:

  • Advice and litigation privilege
  • AI processing potentially privileged content
  • Implications for choice of AI tools

Client confidentiality

  • Professional rules require confidentiality
  • AI tools that store data create considerations
  • Free consumer AI generally inappropriate for confidential matters

Better practices

  • Use enterprise AI with appropriate DPA
  • Specialist legal AI (Harvey, Spellbook) designed with these considerations
  • Local AI for highly sensitive matters
  • De-identify case details when seeking general AI help
  • Document confidentiality practices

Specific tools

For confidential matters:

  • Harvey (enterprise; designed for legal)
  • Spellbook (legal-focused)
  • Microsoft Copilot enterprise (with appropriate DPA)
  • Claude for Enterprise (with DPA)
  • Local AI (Ollama) for maximum confidentiality

Not for confidential client matters:

  • Free consumer ChatGPT
  • Free consumer Claude
  • General free AI tools without enterprise terms

Citation discipline (the most important thing)

If you take one thing from this guide:

Every citation in any document you sign or file must be independently verified.

Process:

  1. AI says case X
  2. Search AustLII for case X
  3. Find actual case
  4. Read it
  5. Verify it supports your point
  6. Cite with confidence

If at any step you can’t verify: don’t cite.

The consequences of submitting fake citations to courts include:

  • Professional discipline (loss of practising certificate possible)
  • Insurance issues
  • Client damage
  • Personal reputation damage
  • Court sanctions
  • Personal cost orders

These outcomes have happened. The technology helps; the discipline must be yours.

A reasonable adoption path

Phase 1: Foundation (Month 1)

  • General AI for non-client work first
  • Internal documents and processes
  • Learning prompt techniques

Phase 2: Contract work (Month 2-3)

  • Spellbook trial or Harvey if firm has it
  • Contract review assistance
  • Verify discipline

Phase 3: Research integration (Month 3-6)

  • AI-augmented research with strict verification
  • Document analysis assistance
  • Communications AI integration

Phase 4: Practice integration (Month 6+)

  • Practice management AI
  • Workflow optimisation
  • Team training and standards

Throughout

  • Citation verification non-negotiable
  • Professional standards apply
  • Stay current on Law Society guidance
  • Maintain confidentiality discipline

Common gotchas

  • Hallucinated citations — most critical risk
  • Confidentiality slippage — free AI tools inappropriate for privileged content
  • Australian law specifics — international AI may use US/UK context
  • Currency — law changes; AI training has cutoffs
  • Statute drafting — AI may produce out-of-date provisions
  • Practice area specifics — generalist AI vs specialist legal AI
  • Indemnity gaps — verify your policy covers AI-assisted practice
  • Court expectations — some courts have specific AI rules

Resources

Australian law sources

  • AustLII (austlii.edu.au) — authoritative free
  • JADE (jade.io)
  • LexisNexis Australia
  • Westlaw AU
  • CCH

Professional bodies

  • Law Council of Australia
  • Law Society of NSW (and state equivalents)
  • Australian Bar Association
  • Harvey (enterprise) — harvey-ai
  • Spellbook — spellbook
  • CoCounsel (Thomson Reuters)
  • Lexis+ AI

Industry media

  • Australian Financial Review legal coverage
  • Lawyers Weekly
  • Australasian Lawyer

A note on access to justice

AI also raises access-to-justice considerations:

  • Self-represented litigants increasingly using AI
  • AI-generated submissions sometimes wrong
  • Court resources strained
  • Pro bono work potentially augmented

For lawyers, this means:

  • Some clients may have AI-prepared materials
  • Quality varies significantly
  • Court patience for AI errors decreasing
  • Opportunity to assist with proper preparation

See also

Sources

  • Law Council of Australia AI position
  • Law Society of NSW AI guidance (2023-2026)
  • Law Institute of Victoria AI resources
  • Queensland Law Society AI considerations
  • Legal Profession Uniform Law
  • Mata v. Avianca (US, 2023) and Australian equivalent cases
  • AustLII, JADE
  • Personal observation of Australian legal AI sector
  • Lawyers Weekly coverage of AI in Australian practice

Graph View

Backlinks