The Client may have a Chatbot, but the Lawyer has an Invoice
From Brisbane to Manhattan to Leeds, lawyers are rewriting engagement terms for clients who bring their own AI. The bill is the least of the clients’ worries.
For three years the profession has fretted about what AI will do to lawyers but perhaps the more pressing question turns out to be what it does to clients.
Australian tax lawyer of Waterhouse Lawyers Tania Waterhouse (above) told the Australian Financial Review about sending an ordinary taxpayer a routine draft of her advice. Late on a Sunday night, it came back stuffed with questions and flowcharts after the client had kindly asked a chatbot what was wrong with it.
And the chatbot, being a chatbot, found plenty.
She adopted a typical legal approach, inserting what she calls an “AI slop clause” in her retainer agreements for those clients who run her work through AI.
The AFR also noted the experience of Queensland tax silk David Marks KC (left) who was recently handed what he called “300 pages of nonsense.” He now reserves the right to pause a matter while he and the client agree on what the cost of assessing that slop may be.
The UK Slop
Britain is heading the same way. A Which? survey found that 51 percent of UK adults now use AI search tools for legal questions, which is a percentage which is only set to increase with a greater uptake on AI by anyone and everyone.
Leeds firm Blacks Solicitors reports more clients arriving with AI-generated documents and advice, and says they bring complications rather than savings. Lawyers there also point out the obvious, that the chatbot accepts no liability for its output, so the client carries any error.
The problem in every jurisdiction, is that lawyers cannot simply bin it.
Professional duties require them to make sure clients understand their advice. Anything that might reach a court has to be verified. Fixed-fee consumer practices take the hardest hit, because no quote ever budgeted for a ten-page chatbot critique.
Not everyone sees disaster in this rush to the chat bots. Jeannie Paterson of the University of Melbourne notes that medicine survived Dr Google, and suggests clients may be less foolish than their lawyers assume.
It’s may be a fair point, and an uncomfortable one for a profession busily marketing AI as the route to faster, cheaper advice. Firms can hardly praise the technology when they use it and condemn it when clients do.
The US Experience
In February, Judge Jed Rakoff of the Southern District of New York ruled in United States v Heppner on documents a defendant had generated, unprompted, with the consumer version of Claude. He held they were protected by neither privilege nor work product.
More than a dozen major US firms have since sent clients warnings. New York’s Sher Tremonte wrote into a client contract that sharing its advice with a chatbot could forfeit privilege.
“We are telling our clients: You should proceed with caution here,” Kobre & Kim’s Alexandria Gutiérrez Swette told Reuters. The law is not yet uniform.
A Michigan magistrate judge treated a self-represented litigant’s ChatGPT-assisted filings as protected work product.
So expect the AI clause to become standard engagement-letter furniture everywhere.
The smart version should say what reviews cost and ask clients to disclose AI use. It should also tell them plainly that their chatbot is not their lawyer and owes them no confidentiality.
It’s all part of the slip-slop world that AI has driven both lawyers and their clients. Billing for the AI intrusion is a very lawyer-focused way to control the slop avalanche.
