Generative AI is no longer “emerging”—it’s everyday
In just three years, generative AI has moved from novelty to normal. The Higher Education Policy Institute (HEPI) / Kortext Student Generative AI Survey 2026 (based on 1,054 UK full-time undergraduates) reports that 95% of students now use AI in at least one way, and 94% use generative AI to help with assessed work. That’s not a trend line—it’s a new baseline.
For K-12 schools, this matters for two reasons. First, today’s students are tomorrow’s post-secondary learners, and the habits they form now will shape how they learn later. Second, many of the same pressures—time, workload, anxiety, loneliness, uneven access to tools, and uncertainty about what’s “allowed”—already exist in school settings.
At TinyEYE, we support schools with online therapy services, and we’re watching the AI conversation expand beyond academics into wellbeing. The survey confirms that shift: a meaningful minority of students are already using AI for companionship and even AI-delivered counselling. Schools don’t need to panic, but they do need a plan.
How students are using AI: less “cheating,” more “scaffolding” (with important caveats)
The survey shows students increasingly use AI for core academic support:
- Explaining concepts (nearly two-thirds report doing this for assessed work)
- Summarising articles (around half)
- Structuring ideas and suggesting research directions (around 40%)
- Searching the internet via AI (36%—a new behavior worth noting)
- Creating media like visuals or audio (13%)
One detail that school leaders should not miss: fewer students report “generating text” than in 2025, but overall AI use is higher. That suggests students may be shifting from one general chatbot to a growing ecosystem of task-specific tools (summarizers, note tools, writing assistants, search copilots, presentation generators). Policies that only mention one tool or one use case will age quickly.
Another critical finding: 12% of students say they have included AI-generated text directly in assessed work (up from 8% in 2025 and 3% in 2024). Even if this is still a minority, the direction is clear—students are testing boundaries, sometimes because expectations are unclear, sometimes because the pressure to perform is high.
Assessment is changing—and so is student anxiety
Nearly two-thirds (65%) of students say assessment has changed significantly in response to AI. Some students report that assignments feel harder, or that formats have shifted. But the most emotionally charged theme is not workload—it’s fear.
Students describe anxiety about false accusations of misconduct, including worries that their writing style could be flagged by AI detection tools even when they did not use AI. This is a crucial insight for schools: when rules are unclear or enforcement feels inconsistent, students don’t just change their behavior—they change how safe they feel.
For K-12 settings, the parallel is straightforward. If students believe they can be “caught” by unreliable signals, they may:
- avoid asking for help
- over-edit their natural voice
- stop taking healthy academic risks
- experience heightened stress around writing tasks
That stress can show up in counseling conversations, classroom behavior, attendance, and motivation—areas where student support teams and therapy providers often get involved.
AI and the student experience: better for many, worse for some
Overall, 49% of students say AI has improved their student experience, while 16% say it has worsened it and 35% report no significant impact. The “better” group often cites:
- time savings
- instant, judgment-free support
- help understanding difficult concepts
- faster progress that frees up time for friends, hobbies, and rest
The “worse” group raises a different set of concerns:
- fairness (some students use AI heavily, others avoid it)
- skill erosion (“I’m not using my brain at all”)
- social isolation (studying alone with AI instead of peers)
- fear about future employment, especially in creative fields
This split is a reminder that AI is not a single intervention with a single outcome. It amplifies what’s already present: strong study strategies can get stronger, but avoidance patterns can also deepen.
Wellbeing and loneliness: the signal schools can’t ignore
One of the most relevant additions in the 2026 survey is its focus on loneliness and wellbeing. The results are nuanced:
- About 41% of students say AI affects their loneliness
- It’s almost evenly split: 21% feel less lonely, 20% feel more lonely
- 15% report using AI for friendship, company, advice, or tackling loneliness
- 8% report using counselling or therapy provided solely by AI, and 4% partly by AI
For schools, this raises practical questions: When a student turns to an AI “companion,” what need are they trying to meet? Is it after-hours support? Fear of judgment? Lack of trusted relationships? Convenience? Or a barrier to accessing human help?
AI can sometimes function like a “study buddy,” but it can also become a substitute for connection. Student support systems should be ready for both possibilities.
Equity: access gaps persist even when tools are “free”
The survey points to an ongoing equality challenge: students from more advantaged households are more likely to use certain AI tools (including text generation and coding support). Importantly, the report notes that this disparity is not fully explained by cost alone, since free versions exist.
In school contexts, this mirrors what we often see with other digital supports: access is shaped by more than subscription fees. It includes:
- device quality and reliability
- internet access and privacy at home
- adult guidance and digital confidence
- time, space, and stress levels that affect whether students can explore tools safely
If schools do not provide structured guidance and equitable access, AI can widen gaps under the banner of “personalized learning.”
What institutions should do next (and what schools can borrow right now)
The HEPI report offers clear recommendations for higher education. Many translate directly to K-12, especially for districts trying to balance innovation with integrity and wellbeing.
- Provide structured AI induction and transition support. The survey notes a third of students arrive without prior AI experience. Schools can build age-appropriate AI literacy so students aren’t left behind—or pushed into risky use.
- Teach AI literacy explicitly, not implicitly. Students need skills in critical evaluation, ethical awareness, and responsible use, alongside subject-specific applications.
- Publish clear, assessment-specific guidance. Uncertainty fuels anxiety. Clear expectations also protect teachers and reduce conflict.
- Ensure equitable access to approved tools. If AI is allowed or expected, access should not depend on a student’s home environment.
- Train staff and protect time for learning. Educators need practical training, not just policy memos.
- Monitor wellbeing impacts, including loneliness. The report calls for research; schools can also track patterns through counseling referrals, student surveys, and family engagement.
A note for student support teams: “AI-only help” is a cue, not a solution
When students use AI for advice, companionship, or even therapy-like interactions, it doesn’t automatically mean harm—but it does suggest an unmet need. This is where school-based supports matter.
TinyEYE’s work with schools is grounded in the idea that support should be accessible, timely, and delivered by qualified professionals. As AI becomes more present in students’ lives, schools can reinforce a simple message: tools can support learning, but students deserve human care for human challenges.
For more information, please follow this link.
