According to AI specialist VerbaFlo and purpose-built student accommodation (PBSA) provider Homes for Students, hundreds of thousands of businesses are at risk of misinterpreting Article 50 of the EU AI Act, with thousands in the property and rental living sector particularly vulnerable.
The new Article, which went live on 2 August, aims to give organisations a working framework for AI transparency and applies to businesses whose AI use falls within its defined scope, such as systems that interact directly with people or generate synthetic content, across Europe.
Given the property sector’s use of AI for initiatives such as chatbots, marketing data procurement and analytics, and its relatively recent adoption of such practices, interpretation is likely to be particularly complex for real estate stakeholders without careful scrutiny of the article and Act.
VerbaFlo and Homes for Students have sought to analyse the framework into guidance and compliance to share best practice on behalf of the industry. They have delivered a joint whitepaper on the requirements for transparent AI deployment under Article 50 of the EU AI Act, the outcome of ongoing collaboration and commitment between the two organisations on responsible AI in residential real estate.
“This whitepaper reflects our own view that AI transparency is a design discipline and should be embedded around the end user experience, rather than a compliance checklist.
“Whilst all businesses need to understand and comply across Europe, the property sector in particular are relatively recent adopters of AI when it comes to its practice and implementation and, we felt that as a leader in these practices, we are responsible to help shape the thinking around it.
“We are pleased to have developed this work with Homes for Students, whose operational scale gave the analysis a level of practical grounding that is difficult to achieve when working solely on the abstractness of theory.
“We hope it gives business teams across the sector a clear, workable reference point which will help them understand what they need to do and avoid non-compliance and potential penalties.”
Sayantan Biswas, Founder and CEO, VerbaFlo
The whitepaper, titled ‘The EU AI Act, Article 50: What Businesses Need to Know Right Now’, reviews the practical implications of Article 50 for operational, marketing and resident engagement functions across the sector, drawing on the companies’ combined experience as an AI provider and a residential operator.
The whitepaper examines three core requirements the companies consider essential to closing the gap between Article 50 compliance in principle and in practice. Article 50 divides responsibility between the organisation that builds an AI system and the organisation that deploys it.
The whitepaper sets out where each responsibility sits in a typical marketing and leasing stack and cautions marketing teams against assuming the AI vendor carries the full weight of compliance simply because the vendor built the underlying system. According to the report, the deploying organisation holds the real responsibility. As a result, marketing and leadership relationships and trust, both internally between divisions and externally with third party suppliers, become essential to get right.
“Article 50 isn’t an abstract legal milestone – it’s a practical line in the sand for how we use AI with our customers. As marketing teams, we don’t just have a compliance obligation, we have a responsibility to be honest about when AI is involved and what it’s producing.
“In student accommodation and Build to Rent that means being clear when applicants are talking to a chatbot, visibly labelling AI-rendered visuals that look like photography, and preserving the integrity of AI-generated content.
“The organisations that treat transparency as part of the customer experience, not as a box-ticking exercise, will be the ones that earn and keep trust as AI becomes embedded in every stage of the leasing journey.”
David Chadderton, Chief Marketing Officer, Homes for Students
The whitepaper argues that disclosure obligations should be assessed against how an AI system is actually experienced by a resident, rather than against its stated function alone. Design choices that make an AI system feel more natural, including voice and persona, generally strengthen rather than reduce the underlying disclosure requirement, and AI generated visual content requires a standard of labelling that goes beyond embedded technical metadata.
The whitepaper also identifies AI used in creditworthiness scoring and guarantor referencing, frequently supplied by third parties, as a category that can sit within the Act’s high-risk framework entirely separately from Article 50. It argues that operators should resolve this classification question well ahead of the framework’s later application date for this, rather than treating it as a future concern.
Overall, the paper unwinds the lower risk activities that marketing teams are already well ensconced with thanks to the practice and compliance of, for example, GDPR, with the higher risk activities of operational teams when it comes to data collection and processing. Most importantly, it showcases where compliance lies, versus framework guidance.
