Tenants are using artificial intelligence to assist with applications to the Tenant Tribunal.
photograph: RNZ
Tenants using artificial intelligence to assist with tenant court applications are creating redundant work and backlogs, according to one property manager.
David Faulkner, general manager of property management at Property Brokers, said there has been a noticeable shift towards AI-created applications, which tend to be longer, more complex and sometimes charge exorbitant amounts of money.
He said his company received a notice of hearing on the lease dispute just before Christmas. One tenant claimed that his and his children’s safety were at risk because he was renting a rural property with unsafe drinking water.
“There were other problems: a retaliation notice, a quiet enjoyment violation, and a broken dryer that took a total of four weeks to repair. Her bill was $40,000.
“What we received was a total of 215 pages consisting of the application, evidence, photographs and a 101 page written report outlining the claim and breakdown of the costs payable to her.The basis of the claim was that the pH level of the water was unsafe for drinking.Two hearings were held, one remotely and once in person.
“On April 2, the judge announced a court order. The tenant was awarded a total of $80 in damages for the inconvenience of the dryer.”
He said such excessively long and seemingly AI-generated claims put pressure on property owners as well as staff.
”[That] This is probably the first one, and it seems very complicated. Then a couple more people start asking for really high amounts of $40,000, $50,00, $60,000, all set in the same format. It was revealed that it was generated by AI.
“Tenants may have a complaint, but they put it together and in the good old days they probably would have gone to something like a tenant union or a tenant advocate and at least had a conversation, but in some cases the AI is just providing unverified information. And they’re probably so excited about what they think they can get that they file it. And that’s starting to cause some problems.”
He said tenancy court judges had to read every page of evidence, which slowed down the process.
Faulkner said some of the claims are baseless and others are outrageous.
Landlords typically bring the majority of applications to the Tenancy Tribunal on the grounds of rent arrears. Faulkner said these cases often do not need to be processed through a hearing, and courts can speed up the speed at which cases are heard by processing them remotely.
Sarina Gibbon, director of Tenancy Advisory, said it’s true that more complex and sometimes frivolous lawsuits are being filed.
But she said this is likely another stage in the evolution of the technology.
Sarina Gibbon, Tenancy Advisory Director;
photograph: Supplied
“If you look at the system as a whole in terms of a common challenge that we’re facing right now, which is making the court pipeline flow more efficiently and minimizing wait times so that genuine applications don’t have to wait months and months and months to get heard, a very rudimentary understanding of AI won’t help you in that sense.
“But if you take a step back and think about the overall intent of the tenant-specific tribunal, this tribunal is actually set up to facilitate cheap and speedy justice and dispute resolution…Would it help if AI enabled more applicants to file applications? Yes, I think it would help. It would serve justice and give more people access to justice.”
“I don’t think there’s anything wrong with AI per se. I think this simply reflects the fact that this technology is so new that we are such naive and primitive users that we don’t yet understand what it means. So what’s happening right now is that a lot of people are working on AI with very little knowledge of RTA itself. So they’re buying into the illusion of AI, and they’re buying into the huge amount of bloat that it creates.”
“I love technology, and I believe that over time, AI will get us to a point where it will be a help rather than a hindrance.”
This comes after the Insurance and Financial Services Ombudsman warned last week that information provided by AI should be fact-checked when people make complaints.
In one example identified in the IFSO scheme, the Google AI brief suggests that claims decisions are “frequently overturned” by consumer complaints, and that “up to 80 to 90 per cent of cases can be successful if people persist.”
But Karen Stevens, from the Insurance and Financial Services Ombudsman, said that was misleading.
She said escalating complaints are often complex and stressful and cannot always be resolved in the consumer’s favor.
He said AI responses can further frustrate people’s processes when reality doesn’t match expectations.
“While AI can be very useful for general information, it can sometimes oversimplify the wording of complex policies, miss key exclusions, or rely on information from overseas that doesn’t apply to New Zealand. We’ve also seen instances where AI hallucinates past events and uses them as examples, which can give the wrong advice,” she said.
“I’ve seen complaints that were as long as 300 pages. But more words aren’t always better. Clear information about what went wrong for someone is much more helpful than multiple pages referencing the law and case law.”
Sign up for Money with Susan Edmundsa weekly newsletter covering everything that affects how we earn, spend and invest money.
