Summary

• Docsity says 91% of surveyed students across six English speaking higher education markets use AI in their studies.
• The survey also found concentration difficulties, home based studying, and video led learning patterns among respondents.
• The business angle is AI literacy, assessment design, graduate readiness, and the edtech market’s shift from content access to learning support.

Docsity says 91% of students in a new international survey are already using AI in their studies, adding to the evidence that higher education has moved from debating whether students will use AI to deciding how its use should be taught, governed, and assessed.

The survey collected responses from 350 higher education students across the UK, US, Australia, Canada, Ireland, and New Zealand. Docsity is a collaborative study and note sharing platform with access to 20m secondary and higher education students across more than 180 countries.

The findings also point to changing study behaviour beyond AI adoption. Docsity says 79% of respondents study from home, 35% feel most productive between 5pm and 9pm, and 41% name difficulty concentrating as their single biggest study challenge. Video and online creators were cited by 39% as the most valuable learning source, ahead of work experience, friends and family, and social media.

The AI figure should be read cautiously because the sample is small and drawn from students likely to be engaged with digital study tools. It is nevertheless consistent with a wider pattern. UK focused research from HEPI and Kortext in March found that 95% of full time UK undergraduates used AI in at least one way.

Students are already using AI for summaries, explanations, organisation, drafting support, revision, flashcards, and concept maps. The harder question is whether they are learning how to use those tools well, and whether universities are redesigning teaching, assessment, and academic integrity policies around that reality.

AI literacy becomes a work skill

Universities now face several pressures at once. Academic integrity policies need to distinguish between legitimate support, poor practice, and misconduct. Assessment design needs to test understanding and judgement rather than only written output. Teaching staff also need support, because students are often adopting tools faster than institutions can provide clear guidance.

Employers will inherit the results. If students use AI throughout their studies, graduate readiness will depend on whether they understand its limits: hallucinations, weak sourcing, over reliance, data privacy, bias, intellectual property, and the difference between assistance and substitution. The labour market will not reward AI use on its own. It will reward people who can verify outputs, apply domain knowledge, and work responsibly with tools that are now routine.

Edtech providers are being pushed into a more active role. The market is shifting from repositories and note sharing platforms towards study companions that summarise, quiz, map concepts, and help organise learning. Docsity’s own AI tools sit in that pattern, turning documents, lectures, photos, videos, and links into study materials.

That creates opportunity and risk. Tools that make studying easier can support access and productivity, especially for students managing heavy workloads, caring responsibilities, neurodiversity, or language barriers. They can also encourage shallow learning if they reward speed over understanding.

The concentration finding adds another layer. Study technology is now being designed for an environment shaped by notifications, short form video, and fragmented attention. Edtech products are competing not only with textbooks or lecture notes, but with the wider attention economy.

Docsity’s survey is a snapshot rather than a definitive measure, but it points in the same direction as larger studies. AI is already embedded in student work. Universities, employers, and education technology companies now have to build practical AI literacy around it, rather than pretending that prohibition can return the sector to a pre generative AI world.