The citations looked legal until the court checked them
The appeal did not rely only on nonexistent case names. It used genuine authorities while changing dates, quotations, and citation details. That mixture is especially dangerous because it gives a reviewer familiar landmarks while hiding invented support between them.
One misquotation concerned a Tennessee landlord's maintenance duties. A source check exposed the difference between a plausible sentence and the law that actually governs the case.
AI amplified an appeal that already had serious defects
The tenant was behind on rent and did not provide the appellate record needed to challenge the lower court's findings. The appeal also leaned on a uniform law that Tennessee has not adopted.
Those failures would not disappear if every citation had been genuine. The lesson is more precise than blaming AI for the judgment: generated legal language can create confidence without supplying the facts, procedure, or jurisdictional fit a case requires.
Access to a chatbot is not access to justice
People facing eviction may have little time, money, or access to counsel. A chatbot can make legal language easier to produce, but it cannot guarantee that the law is current, controlling, accurately quoted, or applicable to the record.
Courts, legal-aid providers, and tool makers should help self-represented litigants verify citations and understand procedural requirements before a polished filing becomes a source of false hope.
- Open every cited authority and confirm the quotation in context.
- Check that the law applies in the correct state, court, and procedural posture.
- Separate factual and record defects from problems in generated language.
- Expand affordable legal review for people facing loss of housing.
Go to the source
Read the evidence behind this analysis. External links open in a new tab.
WKRN — Nashville renter loses eviction appeal after using AI legal arguments Tennessee Courts — Robinson v. Robinson opinion
