Understanding Housing Fraud: How AI Could Change Housing Fraud Detection and Prevention

Ad/Marketing Communication

This legal article/report forms part of my ongoing legal commentary on the use of artificial intelligence within the justice system. It supports my work in teaching, lecturing, and writing about AI and the law and is published to promote my practice. Not legal advice. Not Direct/Public Access. All instructions via clerks at Doughty Street Chambers. This legal article concerns AI Law.

Introduction

I’m grateful to Andrew Lane for collaborating with me on this post. Andy is a Barrister specialising in social housing, local government and public law. He is the author of the “Cornerstone on Social Housing Fraud”. His extensive knowledge and insights have significantly enriched this discussion.

Understanding Tenancy Fraud

Housing fraud typically involves individuals unlawfully subletting their property, providing false information to obtain or succeed to a tenancy, or remaining in a property when they are no longer entitled to do so. In the UK, such fraud is often associated with social housing, although private tenancies are also at risk. According to CIPFA’s Fraud and Corruption Tracker, tenancy fraud is one of the most prevalent forms of fraud faced by local authorities, costing millions of pounds each year and exacerbating housing shortages.

Common Forms of Tenancy Fraud

As already highlighted, there are numerous forms of housing fraud and in our experience new and creative means are demonstrated in court all the time. However, typical examples include:

1. Subletting i.e. a tenant rents out their entire home or part of it without the landlord’s knowledge or permission. Meanwhile, they are living elsewhere, and sometimes have been throughout the entire tenancy.
   
2. False Right-to-Buy Applications. Where individuals misrepresent their satisfaction of the legal requirements needed in seeking to purchase a council house under the Right-to-Buy scheme.
   
3. Allocation misrepresentations. A surprisingly common example of this is where an applicant for social housing fails to disclose that they in fact own or have a tenancy in respect of another property. It also covers those seeking to succeed to a tenancy or to be made a discretionary offer upon the death of a tenant based on false information, such as if and how long they lived with the deceased tenant prior to their death.
   
4. Fraudulent bookings for short-term lets, especially by means of fake third-party ads, websites and telephone calls. Earlier this year Airbnb and Get Safe Online advised on how to avoid fraud and book lettings safely, particularly during busy periods such as Easter.

5 Ways AI May Increase Tenancy Fraud

As people become increasingly familiar with AI tools, related risks are likely to escalate. Five key areas of concern include:

1. Creating Fake and Persuasive Documents Generative AI can produce highly convincing forgeries, such as utility bills, payslips, and reference letters. This also makes it easier for fraudsters to fabricate entirely fictitious identity documents or alter existing ones.
   
2. Utilising Deepfake Voice and Video. AI-driven deepfakes can convincingly imitate a person’s appearance and voice in interviews or online applications. Fraudsters could exploit this technology for virtual property viewings or verification calls, misleading landlords or agents into believing they are communicating with legitimate prospective tenants.
   
3. Automated Bot Attacks on Rental Portals. AI-powered bots can mass-apply for tenancies using multiple fabricated identities or instantly respond to new listings before genuine applicants. These bots can also generate automated references and supporting documentation rapidly, complicating detection efforts for property owners.
   
4. AI-Assisted Social Engineering and Phishing Scams. AI tools could craft personalised phishing emails or messages, convincingly posing as legitimate property portals or agents. This could deceive landlords, tenants, or property management staff into disclosing sensitive financial or personal information.
   
5. AI-Enhanced Financial Fraud. AI algorithms could manipulate rental market listings or pricing information on digital platforms, tricking users into overpaying deposits, rental fees, or administrative charges. Such manipulation could lead to direct financial losses for tenants and reputational damage for genuine landlords or letting agents.

5 Ways AI Can Help Detect and Prevent Tenancy Fraud

Although AI may be used to contribute towards housing fraud, not least in the investigation element, and it may also prove key in developing fraud detection methods including:

1. Advanced Identity Verification. AI-powered ID verification tools can match facial features to identification documents, check digital footprints for consistency, and automatically verify the authenticity of submitted paperwork. Software can flag any discrepancies, such as mismatched personal details across documents.
   
2. Predictive Analytics and Machine Learning. Machine learning models can be trained on data from genuine and fraudulent tenancy applications, spotting subtle indicators of fraud. By analysing patterns, like frequent changes in income details or repeated phone numbers and email addresses, AI solutions can identify applications meriting closer investigation.
   
3. Behavioural Analysis. AI can monitor behavioural patterns on property websites or portals. For instance, if multiple applications come from the same IP address or if user activity shows unusual or repetitive keystrokes, the system can trigger alerts. This approach helps detect whether a human or automated bot is behind the applications.
   
4. Document Forensics. AI-driven image recognition software can detect inconsistencies in document layout, fonts, or tampering traces. Optical character recognition (OCR) can cross-check details in a utility bill (e.g., an electricity provider’s historical usage patterns) to validate the authenticity of claims made by the applicant.
   
5. Link Analysis. Fraudsters often reuse certain details across multiple fraudulent tenancies. AI tools can automatically perform graph or link analysis, showing connections between a suspected individual’s personal data, related accounts, or properties. This can reveal a network of fraudulent tenancies tied to the same group.

Comment

As AI is utilised in the ways outlined above, the legal and investigative processes will likely need to evolve rapidly to respond effectively. Courts will likely see a significant increase in cases involving technologically sophisticated deception, such as deepfakes and AI-generated documents, creating unique evidential challenges. Sticking with the theme of five, we suggest five issues will need to be carefully considered.

1. Evidential norms will need to be revisited. Lawyers and judges must familiarise themselves with AI-generated evidence. Traditional authentication methods, such as physical signatures or contemporaneous documents may require more scrutiny and carry less wait. Instead, digital forensics and AI verification tools could become standard practice within housing law proceedings. This may lead to the courts expecting solicitors and barristers to be proficient not only in law but also in understanding the technological foundations of such evidence.To this end, it is good to see that the Judiciary have had the benefit of ‘Artificial Intelligence (AI) Guidance for Judicial Office Holders’ since 2023.
   
2. Similarly, and following on from this, there may be an increased reliance on expert evidence. Given the complexity of AI-assisted fraud, legal professionals will need to engage specialist digital forensic experts regularly. These experts will help establish the authenticity of evidence presented, evaluate the reliability of AI-driven detection tools, and explain complex technical evidence to the court clearly and persuasively. A related issue is the extent to which experts themselves rely on AI-generated evidence, and how this may impact on their conclusions.
   
3. Case management and pre-action. Courts may need to introduce specific procedural rules requiring advanced disclosure of digital evidence or expert reports validating the authenticity of documents or digital identities. There is some provision for this already in the disclosure rules to be found in CPR 31, not least Practice Direction 31B (disclosure of electronic documents), but does that go far enough and is that too late in the process? Maybe early identification of fraudulent applications using AI tools could become part of pre-action protocols, aimed at reducing the volume of unnecessary litigation.
   
4. Liability and Accountability. AI-driven fraud introduces novel questions around liability. Courts will have to consider where the liability falls when an AI system is deceived. For example, if an AI driven property portal inadvertently facilitates or causes a fraud is there at least a portion of the liability to be shared? What if a landlord fails to deploy adequate AI detection measures when such measures become widely adopted?
   
5. Legislative intervention. Given the potential scale and complexity of AI-driven fraud, courts will undoubtedly strive to ensure justice is served and fraudsters do not profit from wrongdoing. However, judicial intervention alone has limits. Ultimately, Parliament may need to introduce targeted legislation addressing digital and AI-related fraud explicitly, clarifying the duties and liabilities of landlords, letting agents, online platforms, and technology providers.

In short, AI-driven housing fraud is poised not just to affect the way fraud occurs but to fundamentally reshape the UK’s legal practice, procedural norms, and regulatory framework. The legal and housing sectors will have to adapt proactively to ensure it remains capable of delivering fair, efficient, and technologically informed justice.