7 Surprising Truths About How Students Really Use AI

Introduction: Beyond the Hype and Panic

 

For the past couple of years, the conversation around students and artificial intelligence has been dominated by a palpable sense of anxiety. We’ve all heard the headlines and the hallway chatter - fears of widespread, undetectable cheating, the slow erosion of critical thinking, and the looming threat of a "cognitive debt" where students outsource their learning and forget how to think for themselves.

 

But in Series 12 of the AI in Education podcast we spent time listening to a wide range of students, from curious middle schoolers to ambitious university attendees. We felt our job was to tune out the noise and listen to the signal. And the signal is telling a very different story. The reality of how this generation is engaging with AI is far more nuanced, sophisticated, and frankly, more hopeful than the panic suggests. It's a story of active, often brilliant, adaptation. Here are the seven most surprising truths we learned, directly from them.

 

1. The Great "Cheating Panic" Is Largely a Misunderstanding

While it’s true that some students use AI to cheat, the fear of a generation of plagiarists is largely overblown and misinterprets how students are actually engaging with these tools.

Groundbreaking research from Dr. Anna Denejkina [Episode 14] reveals a reality that challenges the common narrative: the vast majority of students, around 80%, state they have not plagiarised using AI and have no intention of doing so.

More importantly, Dr. Denejkina uncovered a crucial, counter-intuitive insight. Many students who think they might be plagiarising are, in fact, using AI for perfectly legitimate learning activities. They're workshopping ideas, brainstorming essay structures, and checking their grammar - processes we would celebrate if they were done with a peer or a tutor.

This is compounded by a very real fear of being falsely accused of cheating, a significant concern noted by Jake Turnbull from Pymble Ladies' College [Episode 8]. The core issue isn't a sudden decline in academic integrity, but a generation left confused and anxious by a lack of clear institutional guidelines. As Dr. Denejkina asked, “It sounds like any use of generative AI for schooling for learning to them is plagiarism? So where have we gone wrong that young people are thinking they're plagiarising when they're actually not?”

 

2. They Want an AI Tutor, Not a Cheat Sheet

Overwhelmingly, students are turning to AI not to bypass their work, but to find a space for immediate, non-judgmental help to better understand it. They are seeking a patient, on-demand tutor that can fill in the gaps left by traditional classroom instruction.

Data from the Chegg survey [Episode 2] was striking: when students struggle academically, 29% turn to generative AI first for support. In contrast, only 8% turn to their professors first. It’s not a rejection of their teachers, but a search for a safe space to be vulnerable. This explains why, in a Harvard Business School case study [Episode 11], a custom AI tutor was primarily used for "concept breakdown" and to ask questions students were "too embarrassed to raise in front of 90 of my peers."

This desire for genuine learning was made explicit at Thomas Blackwood’s school [Episode 8], where students involved in developing a custom AI told the developer in no uncertain terms:

• We want the AI to teach us and not give us the answer.

It’s a sentiment perfectly embodied by 12-year-old Megan [Episode 6], who uses AI for her math homework. She doesn't just ask for the solution; she specifically prompts it to "explain how you did this" and, if necessary, simplify the explanation to a "year one" level. With AI, there is no judgment, only help.

 

3. They're Using AI to Become More Creative, Not Less

One of the most persistent fears is that AI will atrophy creativity, replacing human imagination with machine-generated mediocrity. The story of Caitlin, a Year 11 student, [Episode 10] completely flips that script.

For an English creative writing assignment that required a video component, Caitlin used an AI video generator. Where previous cohorts had relied on the same handful of clips "they found off YouTube or Clickview" that were "just boring," she was able to produce unique, high-quality visuals that brought her story to life.

Crucially, she wrote the entire story herself. The AI wasn't a replacement for her creativity; it was a tool to enhance and visualise it. The process demanded more from her, not less. To get the AI to produce the exact scenes she imagined, she had to engage in an iterative problem-solving process, refining her own descriptive writing when the AI failed. As she explained, "I've had to redescribe it and tell AI... 'I don't want this. That was a bad idea.'" In this case, the AI wasn't a crutch; it was a creative collaborator that demanded a higher level of skill from its human partner.

 

4. They're Becoming Masters of a New Skill: AI Orchestration

Students like Caitlin [Episode 10] aren't just using one AI in isolation; they are developing sophisticated workflows that chain multiple tools together to achieve a complex goal. This is a new, self-taught form of digital literacy: AI orchestration.

Caitlin didn't just type a simple command into the video generator. First, she went to ChatGPT to help her craft a detailed, descriptive text prompt. She then fed that highly refined prompt into a separate AI video generator to get the best possible output. She was using one AI to prompt another.

This is a complex, problem-solving skill that students are developing organically, far ahead of any formal curriculum. They are learning how to manage a team of specialised AI assistants, assigning the right task to the right tool to achieve their vision.

 

5. AI Is Their 24/7 Coach for Academics, Life, and Well-Being

For this generation, AI is becoming a ubiquitous assistant that extends far beyond the classroom. It's an academic coach, a life coach, and a well-being support tool, available 24/7.

The academic coaching is clear, perfectly captured in Brett Moller’s story [Episode 3] of his daughter, who, after getting a math exam back, took a photo of a question she got wrong and prompted her AI, "Please help me, why did I get this wrong?"

But its role as a "life coach" is just as significant. As a report in The Guardian noted [Episode 13], students are using AI for everything from writing internship applications to getting dating advice. The support even extends to mental well-being. Twelve-year-old Megan [Episode 6] shared how she turns to AI when she gets stressed with her dancing:

"I ask it, can you help me calm myself down? And it's like, sure, take a few deep breaths and stuff like that."

For many students, AI is becoming a trusted and versatile first point of contact for a wide range of personal and professional challenges.

 

6. Students Are Actually Teaching the Teachers

In a fascinating power inversion, students are often the most knowledgeable AI experts in the classroom, leading to moments where they are teaching their own teachers.

After Caitlin [Episode 10] demonstrated the stunning videos she had created for her English project, her teacher was so impressed that she wanted to learn how. Caitlin ended up holding an impromptu "master class" for her entire English class, writing the steps on the board for everyone to follow.

This isn't an isolated incident. At Jake Turnbull's school [Episode 8], students ran an entire professional development day for 500 staff members, demonstrating how they use AI in their learning. This role-reversal highlights the incredible pace of technological adoption and underscores the importance of valuing and integrating student expertise into our educational models.

 

7. Clear Rules and Safe Tools Foster Trust and Responsibility

When institutions provide clear guidance and safe, sanctioned tools, students respond with greater trust and more responsible behavior. The confusion that plagues many students is replaced by confidence.

Caitlin's school [Episode 10], for example, uses a "stoplight system" (red for no AI, yellow for polishing and ideas, green for encouraged use). This simple framework removes ambiguity, helps students feel trusted, and empowers them to use AI without fear of accidental wrongdoing.

At Brett Moller's school [Episode 3], student prompts are monitored - not to punish, but to identify opportunities to teach them how to become better, more effective prompters. And at All Hallows' School [Episode12], students noted that they trust the school-provided Gemini account far more than other public tools. The core paradox is clear: providing structure and guardrails doesn’t stifle student use of AI, but rather unleashes it by giving them the confidence to experiment responsibly.

 

Conclusion: Learning from the Learners

If we take the time to listen, it becomes clear that students are using AI in ways that are far more constructive, sophisticated, and hopeful than we often give them credit for. They are not passive consumers waiting for an easy answer. They are strategists like Caitlin, orchestrating multiple AIs to bring a creative vision to life, and lifelong learners like Brett Moller's daughter, turning a moment of failure on an exam into an opportunity for understanding. They are active, critical, and creative users who are navigating a complex new landscape with ingenuity.

Instead of asking how we can stop students from using AI, perhaps the better question is: what can we learn from them about how to use it well?