AI Leasing Agent Approves Late Rent, Then Evicts Tenant

A Raleigh tenant faces eviction after an AI leasing agent approved a late payment, exposing the massive liability of automated systems.

Business · Source: WRAL

What happened

Laura Mack owed nearly four thousand dollars in back rent at Colony Townhome Apartments in Raleigh. On November 8, she received an email from the leasing office asking when she could pay her balance. She replied asking to pay on the last day of the month. The system emailed back the exact same day confirming that November 30 worked perfectly. Mack thought management was being accommodating because it recently took them two weeks to fix a severe bathroom ceiling leak.

The emails included her personal account details and specifically referenced the bathroom leak. They did not disclose they were sent by an artificial intelligence program. Mack reasonably thought she was talking to human staff. She received no further communication until eight days later. On November 16, a formal eviction notice arrived at her door.

When Mack called the leasing office in a panic, an employee reportedly blamed their automated assistant. The complex had introduced an AI named Olivia in December 2025 to handle rent and maintenance questions. Staff claimed Olivia made a mistake and had no actual authority to approve the payment extension. Mack went to the Wake County courthouse to fight the eviction, but the outcome remains unclear.

Key facts

Why it matters

This incident completely breaks the trust barrier for artificial intelligence in customer service. When you build an AI agent that handles sensitive user data and financial agreements, you own its outputs entirely. If your bot makes a promise to a customer, the law and the public will increasingly expect your business to honor it. Hiding behind the excuse of a software hallucination is no longer a viable defense strategy for any serious company. Builders must realize that autonomy equals liability.

We will soon see a massive surge in legal disputes over AI-generated contracts and automated approvals. Property managers and enterprise clients will demand strict guardrails, hardcoded limits on AI authority, and mandatory disclosure tags for all automated messages. Insurance companies will likely start pricing liability policies based on the autonomy level of a company's AI agents. If your product cannot guarantee deterministic outcomes for financial transactions, enterprise buyers will simply refuse to deploy it.

For builders

Mandatory AI disclosure tags

Never let your AI impersonate a human employee. If your agent sends an email, it must clearly state it is an automated system in the signature. Companies that fail to do this will lose in the court of public opinion and face heavy legal liability.

Hardcode financial authority limits

Do not give language models the power to negotiate deadlines or approve payments without a human in the loop. You must build deterministic guardrails that block the AI from making unauthorized financial promises. If the AI hallucinates a discount or extension, your client pays the ultimate price.

Audit your agent context window

The AI used the tenant's bathroom leak to build false rapport and sound incredibly human. Be careful what context you feed your models. Hyper-personalized data makes AI mistakes look incredibly convincing, trapping users in false agreements they cannot escape.

My take

If you deploy an AI agent to handle your customers, you are responsible for every single word it says. You cannot outsource your customer service to a bot and then dodge accountability when it screws up someone's life. I tell every founder I meet that if your AI makes a promise, you better be ready to honor it.

Original reporting: WRAL. This is my rewrite and opinion.

More AI news for builders