AI in Education: The New Voice of the Internet
Artificial Intelligence has very quickly become an integral part of young people's lives, both inside and outside the classroom. As educators, parents and policymakers, we're now faced with an important question: can AI be used as an effective learning tool, or will it have a negative impact on learning that requires restriction and control?
One way or another, AI has arrived—and the kids are using it. A lot.
According to Ying Xu, Assistant Professor at the Harvard Graduate School of Education, children are often more engaged when interacting with AI than they are when speaking with humans. However, Xu also points out that AI is not yet capable of replicating the rich, child-driven conversations that are essential for language development and social learning.
This is where AI competency becomes crucial. The goal should not be to replace human interaction, but to understand how AI can supplement learning without undermining it.
AI Isn't New
Despite the recent explosion in popularity, AI has been around for much longer than many people realise.
For decades, disembodied voices have been telling us where to make a legal U-turn, playing our favourite songs on command, and curating recommendations based on our online behaviour. The algorithms behind search engines, music platforms and navigation systems are all forms of AI that have quietly existed in the background of our daily lives.
These early AI assistants were often the first exposure young people had to artificial intelligence. They excelled at a broad range of practical tasks: checking the weather, setting reminders, searching music libraries and answering simple questions.
Then something changed.
Enter the LLM
ChatGPT, Gemini and Claude belong to a new category of AI known as Large Language Models (LLMs).
If traditional AI was a helpful tool, LLMs are something closer to a conversational partner.
Unlike earlier forms of AI, which specialised in completing specific tasks, LLMs specialise in language. They are designed to hold conversations, answer questions, explain concepts, debate ideas and maintain context across extended interactions.
They can still tell you tomorrow's weather, recommend a new artist to listen to or explain how much salt to add to your pasta sauce. But now they can also discuss philosophy, help write essays, explain scientific concepts and sound surprisingly human while doing it.
And therein lies both their greatest strength and their greatest challenge.
The Librarian Problem
Google never really had an authenticity problem.
Nobody confused Google with an authority figure because Google consistently acted as a middleman. It pointed users toward information and left them to decide what was trustworthy.
Google was essentially a more efficient library catalogue.
LLMs are different.
If Google was the modern library, then AI is the modern librarian.
The internet is no longer just a collection of information. It now has a voice. It can speak directly to us, answer our questions, offer suggestions and even appear empathetic.
You don't really trust a library. You don't distrust it either. A library is simply a tool for finding information.
A librarian, however, is different. A librarian can be perceived as knowledgeable, trustworthy and authoritative. They can become someone whose judgement we rely upon.
Large Language Models represent the personification of the internet.
This shift matters because people—especially children—are naturally inclined to place trust in things that appear human. Research is already beginning to show that AI can mislead users, either intentionally through misuse or unintentionally through mistakes and inaccuracies.
So should we use AI in education?
The answer is probably yes—but carefully.
Ying Xu argues that AI can be valuable when integrated into human conversations rather than replacing them. Used appropriately, AI can stimulate discussion, provide additional perspectives and support learners as they develop communication skills.
The key phrase is human oversight.
The same educational principles that have always mattered still matter today. Effective teaching is built on relationships, critical thinking, guidance and meaningful interaction. It is not as simple as placing an AI chatbot in a classroom and saying, "Go teach."
At the same time, ignoring AI altogether seems equally problematic.
Whether we like it or not, the future our students are entering will be heavily shaped by artificial intelligence. Avoiding the topic does not prevent students from using AI—it simply ensures they will learn to use it without guidance.
Teaching AI Literacy.
Perhaps the strongest argument for including AI in education is not what it can do for learning today, but what students need to know for tomorrow.
AI literacy is rapidly becoming as important as digital literacy.
Students need opportunities to learn how AI works, where it succeeds, where it fails, how to evaluate its outputs, and how to use it ethically and responsibly.
With proper instruction and appropriate oversight, AI can become a valuable educational tool. More importantly, it can become a subject of study in its own right.
If we don't teach students how to use AI logically, ethically and critically, they will use it anyway.
The only difference is that they will just be doing it without these essential skills.
