The Technology Gap the Student Accommodation Sector Isn't Talking About

Half of Chinese students use AI for almost everything. Four in ten American students don’t use it at all. British students are somewhere in between. Our survey of 14,000 residents across 191 properties reveals a divide that’s reshaping how students discover, evaluate, and experience their accommodation, and most operators have no idea it exists.

There is a number buried in our latest student satisfaction survey that has been bothering me for weeks, not because it’s surprising, but because of what it implies about how differently we should be thinking about the students we house.

The Investor in Students Spring 2026 survey collected 14,189 responses across 191 properties and 56 cities. It is, as far as I’m aware, the largest proprietary dataset on student accommodation satisfaction in the UK, and like any large dataset, it rewards close attention. One of the questions we ask is about technology use, including how students interact with artificial intelligence tools in their daily lives. When you split the responses by nationality, the picture that emerges is not a gentle gradient. It is a cliff.

Half of Chinese students reported using AI for almost everything. Not occasionally. Not for specific tasks. For almost everything: researching, writing, planning, communicating, organising their lives, navigating a new country, making decisions about where to live and what to study. AI is not a tool they reach for when they need help. It is the environment they operate in.

Among American students, by contrast, four out of ten reported that they don’t use AI at all. Not rarely. Not for limited purposes. They simply don’t use it.

The UK domestic students sit somewhere in between, though closer to the American end of the spectrum than most people would assume. Recent data from the Higher Education Policy Institute shows that 92% of UK students were using AI in some form by 2025, a sharp increase from 66% the previous year, and 88% had used generative AI for assessments. A YouGov survey in September 2025 found that two-thirds of UK undergraduates were using AI for degree-related work, with ChatGPT the dominant tool at 74% adoption. These are high numbers. They suggest a student population that is rapidly embracing AI.

But adoption and integration are not the same thing.

The Difference Between Using AI and Living in It

The HEPI data tells us that 92% of UK students have tried AI. The IIS data tells us something different about the Chinese students in our properties: they haven’t just tried it. They’ve built their lives around it.

This distinction matters more than the headline numbers suggest. A UK student who uses ChatGPT to help structure an essay is using AI. A Chinese student who uses DeepSeek to research accommodation options, translate between Mandarin and English, plan their weekly schedule, prepare for job interviews, navigate visa paperwork, and maintain contact with family back home is living in an entirely different technological reality. The first is a tool. The second is infrastructure.

The data from China supports this reading at scale. By December 2025, China had 602 million generative AI users, a 141.7% increase from the end of 2024. A 2025 survey by the Tencent Research Institute found that 96% of highly educated young Chinese adults had used AI-generated content tools, with two-thirds (67.7%) using them daily. Among Chinese university applicants in 2026, 78.4% had used AI to search for universities or generate application plans, and over 60% of parents considered AI-generated recommendations a key reference in their child’s decision.

These are not niche adoption figures. They describe a population for which AI is as normalised as search engines were for the previous generation. The Chinese students arriving at UK PBSA properties in September 2026 have grown up in an environment where DeepSeek, ByteDance’s Doubao, and Tencent Yuanbao are as ubiquitous as Google and WhatsApp are in the UK. They don’t think of AI as novel. They think of it as infrastructure, the layer through which information is accessed, decisions are made, and daily life is organised.

The UK students arriving at the same properties, by contrast, are in the midst of a rapid but incomplete adoption curve. The numbers are growing fast, but the depth of integration lags significantly behind. Most UK students use AI for specific academic tasks. Fewer use it as a general-purpose life tool. The technology is present but not yet embedded in the same way.

And then there are the American students, of whom roughly 40% don’t use AI at all. This isn’t a failure of awareness. It’s a different relationship with the technology, shaped by different cultural attitudes to automation, different concerns about privacy and authenticity, and a higher education system that has been slower to integrate AI into its pedagogy and assessment methods.

What This Means for How You Run a Property

The PBSA sector has spent the last decade building a service model around a broadly homogeneous technological assumption: students use smartphones, they use social media, they use search engines, and they expect digital-first communication. That assumption held because the gap between the most and least digitally fluent residents was manageable. A WhatsApp group worked for everyone. A website with clear information served everyone. A reception team that could answer questions in English was sufficient.

The AI divide shatters that assumption.

Consider what it means to house a Chinese student who uses AI for everything alongside a British student who uses it mainly for essays, and to provide them with the same information architecture, the same communication channels, and the same support services. The Chinese student is accustomed to receiving synthesised, contextualised, personalised information on demand. They ask a question and get an answer, not a link to a webpage where they might find the answer if they look hard enough. They expect their accommodation provider to operate at the same level of immediacy and personalisation that their AI tools provide.

The British student, by and large, does not have that expectation. They’re still operating in a search-and-browse paradigm, where information is found rather than delivered, where the website is the primary source of truth, and where a phone call to reception is a reasonable way to resolve an issue. These are not better or worse expectations. They’re different expectations, shaped by different technological realities, and the property that serves both populations identically is serving at least one of them poorly.

The American student who doesn’t use AI at all presents a third set of needs. They may be sceptical of automated responses, wary of chatbots, and resistant to digital-first service models. They want a human. Providing them with an AI-powered FAQ and expecting satisfaction is as misguided as providing the Chinese student with nothing but a static website and expecting the same.

The Tools Are Different Too

There is a dimension to this divide that most Western operators are completely blind to, and it concerns not just how much students use AI but which AI they use.

Chinese students overwhelmingly use Chinese-language AI tools: DeepSeek, Doubao, Tencent Yuanbao, MiniMax, and a constellation of smaller platforms that are deeply integrated into the Chinese digital ecosystem. These tools are not just translations of Western products. They are built for Chinese-language contexts, trained on Chinese data, optimised for Chinese cultural norms, and embedded in the Chinese app ecosystem in ways that ChatGPT and Perplexity are not.

This means that when a Chinese student asks their AI tool “what’s the best student accommodation in Manchester,” they’re getting an answer generated by a system with a fundamentally different knowledge base than the one a British student is querying. The Chinese tool may draw on Chinese-language forums, WeChat discussions, Xiaohongshu reviews, and Mandarin-language university guides. The British student’s ChatGPT draws on English-language sources, Western review platforms, and the English-language web.

The two students are asking the same question and receiving answers constructed from entirely different evidence ecosystems. The PBSA operator who invests in English-language SEO, English-language review management, and English-language PR is visible to the British student’s AI and invisible to the Chinese student’s AI. The operator who builds a presence on Xiaohongshu, WeChat, and Chinese-language review platforms is visible to the Chinese student and invisible to the British one.

Very few operators are doing both. Most are doing neither effectively.

The Cognitive Offloading Problem

There is a longer-term concern that the sector should be aware of, even if it’s not yet a operational problem.

A longitudinal study published in June 2026, tracking 26,811 Chinese secondary school students over 30 months, found that students using generative AI saw an 18% rise in homework scores and a 30% reduction in completion time. On the surface, this is exactly what you’d want: faster, better work. But the same study found that these students experienced a 20% drop in closed-book exam scores within six months, and entrance exam scores declined by 18% to 24% after approximately two years of AI use.

The researchers identified the mechanism: approximately 80% of student AI users demonstrated “cognitive offloading,” the outsourcing of thinking to external tools. The students weren’t learning to think better. They were learning to delegate thinking more efficiently. The short-term performance gains masked a longer-term erosion of independent cognitive capability.

For PBSA operators, this has a direct implication. The Chinese students arriving at your properties are, on average, more AI-fluent and more AI-dependent than their British counterparts. They are better at using the tools and more reliant on them. This creates a support paradox: the students who are most comfortable with AI are also the ones whose independent problem-solving skills may be most eroded by it. When the AI can’t help them, when the question is about a specific property policy, a maintenance request, a social situation, or a mental health concern, the gap between their AI-assisted capability and their independent capability may be wider than it appears.

This isn’t a criticism of Chinese students. It’s a description of a systemic effect that any heavy AI user experiences, and it has practical implications for how properties design their support services. The assumption that digitally fluent residents need less support is wrong. They need different support, support that meets them where they operate but also provides the human connection and contextual understanding that AI tools cannot.

What the Data Demands You Do

The IIS data doesn’t just describe a divide. It demands a response, and the response needs to be more nuanced than “add a chatbot” or “translate the website into Mandarin.”

Segment your digital strategy by nationality. Not as a nice-to-have. As a necessity. The information architecture that serves a British student is not the same one that serves a Chinese student. The channels are different, the tools are different, the expectations are different, and the evidence ecosystems that feed into AI recommendations are different. A single, English-language, search-optimised website is not a digital strategy. It’s a monolingual artefact of a pre-AI world.

Build presence in the Chinese digital ecosystem. Xiaohongshu, WeChat, Chinese-language review platforms, Mandarin-language university guides. If your property doesn’t appear when a Chinese student asks DeepSeek or Doubao for recommendations, you don’t exist for that student. This is not a marketing nicety. It’s a survival requirement for any operator that houses Chinese residents in meaningful numbers.

Design support services for three technological realities, not one. The AI-native Chinese student who expects instant, synthesised, personalised information. The rapidly-adopting British student who is moving toward AI integration but still operates primarily in a search-and-browse paradigm. The sceptical American student who wants a human being, not a chatbot. One service model cannot serve all three. The operators that recognise this and build accordingly will see it in their satisfaction scores. The ones that don’t will see it in their attrition rates.

Prepare for the cognitive offloading wave. The students arriving now are more AI-dependent than any cohort before them. This will accelerate. The properties that invest in human support infrastructure, not as a fallback for when technology fails, but as a deliberate complement to AI-driven information delivery, will be building for the world that’s coming rather than the one that’s already gone.

The Uncomfortable Conclusion

We are housing students from fundamentally different technological civilisations in the same buildings, running the same events, providing the same support services, and marketing to them through the same channels. The IIS data makes the scale of this difference measurable. Half of Chinese students use AI for everything. Four in ten Americans don’t use it at all. British students are somewhere in between, adopting fast but integrating slowly.

The operators that treat this as a translation problem, “we need a Mandarin website,” are missing the point. The divide is not linguistic. It’s cognitive, behavioural, and infrastructural. The Chinese student who arrives at your property in September 2026 is operating in a different digital reality than the British student in the next room. They discover information differently, make decisions differently, evaluate quality differently, and expect service differently. Treating them the same is not fairness. It’s failure.

The data is clear. The question is whether the sector is ready to act on it.