There’s a new participant entering tenancy deposit disputes, and it’s not a landlord, tenant, inventory clerk or adjudicator.
It’s Artificial Intelligence!
Property managers (and adjudicators) are increasingly receiving detailed challenges to deposit deductions that appear professionally written, expertly structured and packed with references to “fair wear and tear”, “burden of proof” and “tenant rights”. In many cases, these responses have been generated with the help of ChatGPT or similar AI tools.
At first glance, this might seem concerning. Has AI suddenly become an expert in tenancy deposit disputes?
Not quite.
The problem with AI analysis
Artificial intelligence is exceptionally good at analysing information that it is given. The challenge is that it has no way of knowing what information it hasn’t been shown.
A tenant may upload a check out report and ask:
“Can my landlord charge me for this?”
The AI will review the report and provide an answer based solely on the information presented. What it usually won’t have access to is:
+ The signed check in report
+ The tenancy agreement
+ The landlord’s invoices
+ Full photographic evidence
+ Previous correspondence
+ Deposit scheme guidance
+ The age and quality of the item being claimed for
Without this context, the AI is often making assumptions rather than reaching evidence-based conclusions.
The missing piece – comparison
An experienced inventory professional knows that a comprehensive check out records change in condition and cleanliness by comparing the move in inventory, but it does not include the full details from that inventory. AI cannot see the full picture, so it cannot provide a correct appraisal.
The real assessment comes from comparing, the Check In Condition → Tenancy Duration →Property Visit and Check Out Condition.
Furthermore, it has advised tenants that they cannot be charged the full cost for a new carpet. However, if the carpet was of high quality, new condition and the large stain in the middle of the room was unable to be cleaned, then a full cost (less depreciation) can be recoverable.
Why cleaning disputes are particularly vulnerable
Cleaning deductions are perhaps the area most likely to generate AI-assisted challenges.
A tenant may upload photographs showing a property that appears generally clean and ask whether a deduction is justified. The AI may respond that the property appears to be in good condition. However, inventory professionals and adjudicators frequently identify issues that are not immediately obvious in a general photograph, including:
+ Grease deposits inside ovens
+ Dirty extractor filters
+ Limescale build-up
+ Mould in washing machine seals
+ Dust accumulation on skirting boards
+ Food residue in cupboards and drawers
A property can appear clean to a casual observer while still falling below the standard recorded at check-in.
The false confidence trap
Perhaps the biggest danger isn’t that AI gets everything wrong. It’s that it sounds convincing.
Tenants may receive a detailed response stating: “The landlord may struggle to justify this deduction” or “This appears to fall within fair wear and tear.”
The language is persuasive and authoritative, yet deposit adjudicators do not make decisions based on persuasive wording. They make decisions based on evidence.
A well-written argument supported by incomplete information is still an incomplete argument.
What property managers should do
The arrival of AI-generated disputes doesn’t mean property managers need to become AI experts. Instead, it highlights the importance of returning to the fundamentals.
One of the biggest mistakes a property manager can make is attempting to rebut every point raised by ChatGPT, so remember to Lead with Evidence, Not Opinion. AI-generated arguments often rely on general principles so property managers should respond with specifics such as “The check in report records the carpet as professionally cleaned. The check out report records six separate stains not present at the start of the tenancy. Supporting photographs are attached.”
Where possible, present evidence chronologically including check in photographs or comments, relevant tenancy clauses, check out photographs or comments, contractor invoices and any supporting correspondence such as a property visit. The more complete the story, the harder it becomes to challenge the deduction on procedural grounds.
Also remember the written word is the primary evidence with photos acting as supporting evidence and not the other way around (ChatGPT doesn’t know this). As long as it’s recorded, additional photos are not always a necessity.
AI-generated disputes can sometimes feel confrontational because they are lengthy and highly detailed so don’t be drawn into defending the principle of the deduction.
When responding to a tenant’s challenge, write every email as though it will be reviewed by an adjudicator later. Professional, evidence-based responses tend to carry significantly more weight than lengthy debates.
The opportunity for professional inventory management
Far from reducing the importance of inventories, AI may actually increase this.
The stronger the inventory process, the less room there is for speculation.
Detailed descriptions, high-quality photographs, signed reports and robust comparative evidence leave little opportunity for generic AI-generated arguments to gain traction.
In many ways, the rise of artificial intelligence reinforces a truth that inventory professionals have known for years:
The strongest defence against a dispute has never been a better argument. It has always been better evidence.
Justine Tomlinson is chief operating officer at No Letting Go.
