Higher education

Why these institutional decisions are failing students, twice.

01 May 2026 / Insights

And what change looks like.

The 2026 Studiosity-YouGov Global Wellbeing Survey reveals a student cohort stuck between two decisions by their universities / institutions:

1. pressure to adopt general-purpose technology, and

2. persisting with a culture of suspicion.

Executive Summary:

- - HE institutions are facing legal, financial, and reputational fallout.
  - ...but easily the most immediate and devastating loss is the erosion of cognitive agency and emotional wellbeing amongst students, and more broadly, the impact to trust in teaching and learning environments.
  - In the latest Studiosity & YouGov Global Student Wellbeing Survey of 10,330 university students, feedback highlights students' concerns, advice to leaders, and suggestions for change.
  - An analysis of these findings in the context of wider discourse, points to three institutional decisions causing frustration amongst students: technology before pedagogy, detection before agency, and product over process.
  - Students make clear their concerns for their own thinking skills, trust in the teaching and learning environment, and mental health - with concern even more visible for the most vulnerable, including younger undergraduates, academically at-risk students, and non-native English speakers.

1. The cognitive failure; adoption without agency

While 81% of students have adopted general-purpose AI, only 41% feel highly confident they are actually learning while using it.

Many students, particularly those at greatest risk, are keenly aware of this trade off:

- - Universal adoption, low confidence: While 81% of students use AI, most do not have high confidence that they are learning while using these tools. This gap in the survey is significantly greater for younger students, and is a persistent finding elsewhere.
  - 'Get it done' is a goal: 37% of students admit their primary reason for using AI is to get university work done, faster. The compromise looks like cognitive surrender, which students recognise as...
  - ...Skill erosion: 44% of AI-using students say that it is actively reducing their critical thinking and communication skills.
  - Confidence and use varies, notably in cohorts under more pressure: 89% of students whose first language is not English use general AI to complete their university studies, compared to 80% of English-first speakers. Further, English-as-first-language students are more confident (43% vs. 36%) they are "learning" while using general LLMs. Meanwhile, First Year students are 13 pp less confident than postgraduate level (35% vs 48%). A problem that is almost certainly likened to handing over the car keys too early.

Most of what I know about using AI responsibly comes from general advice online rather than formal university guidance... — UK.

AI means it’s easier to get the answer without thinking — Australia

It lacks originality and mostly monotonous also feels like cheating when using it to do something that requires critical thinking — Saudi Arabia

I use it to brainstorm and feel like I'm not using my brain at all — Australia

Because it makes me lazy to search, learn in traditional way, which will allow me to read more to filter out which is right which is wrong — Malaysia

Using AI can sometimes shortcut the thinking process, which is important for building strong critical thinking and communication skills. — Australia

I have read articles and watched videos that claim that studies show people’s critical thinking skills have gone down with the use of AI, even though I may not feel like it has happened to me, that I might just be unaware — Canada

In a sense i feel babied into knowing things or finding the answer instead of having to skills to figure it out myself. i also worry that it may lead to me relying on ai for answers in the future — Australia

When using ai it feels wrong but when there isn’t support it feels like the only thing I can use to help me. — Australia

I use AI to generate ideas sometimes and that makes me worry that I won't ever be able to form my own ideas or even thoughts without AI. — Australia

I think it helps me complete an assignment quicker but I sometimes worry if I wouldn't have been able to complete the assignment to the same level of work if I didn't use AI — Canada

Good for clearing doubts but fear of getting too addicted — Canada

2. The affective failure; a climate of suspicion in T&L

Students are using the consumer-driven AI tools they are told are essential for their graduation and professional prospects...

...yet must do so in a policed environment that creates an affective drain.

- Fear of false accusation: 77% of students who use AI report anxiety about being "wrongly flagged" or falsely accused of cheating by detection tools.
- AI addiction and guilt: In markets like Malaysia (55%) and the UAE (46%), students cite "being addicted to using AI" as a primary source of stress.
- Uncertain, or no guardrails: 42% of the 10,330 students had no knowledge of training, policies, or guidelines from their institution around proper use of LLMs.
- The human connection void: As students navigate an ethical minefield, human connection is vanishing; only 40% of students globally had access to a mentor in 2026, an 18 percentage point drop year-over-year. And university is nothing if not a social experience.

Closer to final year, I'm getting more stressed and pressured. — Malaysia.

I don’t use AI for this purpose. But I have heard of students who also don’t use AI, but whose work has been flagged. That’s why I feel stressed about it - just because I don’t use it doesn’t mean to say that my work might not be (incorrectly) flagged. — United Kingdom

There’s so many people who got flagged by the AI bot even though they didn’t use AI. — Singapore

Since we are using AI in every single aspect and the detection software getting stronger and better. Its inevitable that everything ever written will be detected by the anti AI system. — Malaysia

AI makes me nervous. — Australia

I've tended to avoid certain phrases, words and persuasive techniques due to their new association with AI language models, limiting the range I can express myself with without fearing getting flagged as using AI to write my work. — United Kingdom

3. The result, students and degrees under pressure.

The useful struggle essential to learning is being replaced by a different kind of - less useful struggle - frustration and fear.

From students' feedback, we can see this sentiment appears to be directed two ways:

1. intrinsic - self, feelings of guilt or uncertainty; and

2. extrinsic - negative perceptions of their the teaching and learning environment.

Learning feels unsafe.

Perhaps little wonder that there is also a decline in student optimism upon graduation:

1. Job confidence: Confidence in securing a job related to their degree within six months of graduation has dropped from 59% in 2025 to 55% in 2026.

2. Career prospects: 41% of students globally fear AI will make it harder to get a job in their field, a concern that peaks as high as 48% in Singapore.

4. The way forward looks like educators leading the return to pedagogy, first.

Higher education faces a dual crisis at the core of discussion right now:

- - a technology-led pressure to prioritise AI Literacy at the expense of learning;
  - a resulting tension that leaves students feeling guilty, under surveillance, conflicted, fearful, and uncertain about their teaching and learning environment.

To resolve this, universities and leading voices are bringing back their pedagogical 'filter' - not only for what is put in front of students, but to defend the holistic teaching and learning system around them.

This approach will look like: guaranteeing degree value for students (and their families). Specifically, that a branded degree journey will prove the development of persistent, critical thinking skills and other 'owned' skills that are indispensable for life chances.

More than 10,000 university students left comments in the survey this year, tasking their institutional leaders with higher standards, with recurring requests:

1. Guardrails: A clear institutional position on cognitive offloading, that eliminates fear of being wrongly flagged for misconduct, and removes frustration from students who want to learn - but are academically penalised for opting out of general LLM use.
2. Career relevance: Integration of AI tools into the curriculum to ensure any formalised AI use is of recognised skills for their degree - and specifically, to remain competitive in a job market where they feel threatened.

3. Wellbeing support and human connection: Specialised human connection and counselling to address the unique mental health pressures of the digital-first era.

Fortunately, educators are some of the loudest advocates for change.

And when given the opportunity, we can see that students are too.