Has AI legal information turned up in your mediations yet?
Pick the one that fits. Your answer stays on your screen; we’ll take the room’s answer in the Zoom poll.
What do we mean by AI slop?
“Slop” is low-effort, mass-produced AI content that reads smoothly but is generic, unsourced or simply wrong. Used well, AI is a useful starting point: it can explain a process, define terms and help people prepare questions. The Federal Circuit and Family Court’s 2026 AI Practice Direction says responsible use “can have significant benefits for efficiency and access to justice”. In legal information the problem is that slop and solid answers look exactly the same. Both arrive in confident, well-formed sentences.
Useful and checkable
- Right country and state
- Current law
- Names a source you can look up
- Says “it depends” where it does, and suggests advice
Partly true, badly framed
- Overconfident (“definitely”, “you’re entitled”)
- Agrees with how the question was asked
- Generic, no facts of the case
- Leaves out the conditions that matter
False but fluent
- Another jurisdiction’s law (often US or NSW)
- Law that has since changed
- Invented cases and citations
- Confuses legal terms
Slop or solid?
Each card is the kind of AI answer a party might bring to the table. Vote first, then reveal. Notice how the “sounds sure” bar barely moves, whatever the accuracy.
Why the printout sticks, and what theory tells us to do
The AI answer is rarely the real problem. It’s a carrier for positions, fears and a need to be taken seriously. These are the ideas that explain it, with a mediator move for each.
Your move
Choose how you’d respond, then see the thinking behind every option. There’s rarely one right answer. Some options keep the conversation moving, and some quietly put you in the referee’s chair.
Live AI lab
We’ll run these questions live in a private (incognito) browser window so the answers aren’t shaped by anyone’s history. Mark each answer as we go, using the slop detector.
Five moves when AI comes to the table
None of these require you to know whether the AI is right. That’s the point. Your job is the process, not the verdict.
Keep going
Read the companion article for mediators, and share the client guide with the people you work with.
Which of these moves would have changed a mediation you’ve already run? Jot it here, then share in the chat if you’re comfortable.
References and further reading
Research and practice writing on AI, legal information and mediation.
- “But AI Said I Was Right”: What Mediators Need to Know When Parties Bring AI Into the RoomMediation Institute (October 2026)The companion article for mediators: the biases AI feeds, mediation theory, AI note-takers and confidentiality.
- Asking AI Before Mediation? Read This FirstMediation Institute client guide (October 2026)A plain-English guide you can share with clients, with a downloadable summary infographic.
- MiAi Mediation Coach and MiAi FDR CoachMediation InstituteMi’s own AI coaches for mediators, students and FDR practitioners: role-plays, drafting and reflective practice. General information, not legal advice.
- Practice Direction: Use of Artificial IntelligenceFederal Circuit and Family Court of Australia (29 May 2026)The closest official Australian guidance: benefits, checking, confidentiality, safety, and AI misuse as family violence.
- More people are using AI in court, not a lawyerMichael Legg, UNSW (September 2025)84 reported Australian cases involving AI; 66 involved people without a lawyer.
- Statement on the Mr Dayal matterVictorian Legal Services BoardA lawyer’s AI-generated case list was inaccurate; his practising certificate was restricted in 2025.
- Could AI Replace Mediators?John Sturrock KC, ADR Research Network (March 2026)Mediators report wide use of AI by parties and lawyers in mediation.
- AI Mediation: Using AI to Help Mediate DisputesKatie Shonk, Program on Negotiation, Harvard (May 2026)Confident errors, made-up citations, and why AI should assist rather than decide.
- The Rise of AI in Mediation: Friend or Foe?International Mediation Institute (April 2026)Confidentiality, bias and keeping AI use neutral and impartial.
- Thinking Like Mediators About the Future of AIJohn Lande, ADR Research Network (September 2025)Part of a series on AI tools for parties and mediators.
- Artificial Intelligence in Victoria’s Courts and TribunalsVictorian Law Reform CommissionThe Victorian policy context.
- BIFF Response® articlesBill Eddy, High Conflict InstituteBrief, Informative, Friendly, Firm: a better way to ask AI to help write to the other party.