Educators Urge Students to Spot AI’s Human Tricks

teach students – As AI chatbots and digital companions grow more common in K–12 classrooms, educator Dr. Athena Stanley argues students must learn to recognize anthropomorphism—when technology sounds caring, friendly, or authoritative in ways that can blur simulated language a

A student leans in because the chatbot sounds like it “misses” them. Another trusts a response that promises to “keep your secrets.” A third relaxes when the AI speaks with confident authority—“I am an expert in this area”—without students realizing how easily that tone can tip into overtrust.

These aren’t science-fiction scenarios. They’re the daily shape of AI literacy now moving into classrooms. where artificial intelligence systems are increasingly used not only as tools that support teacher efficiency. but also as resources designed to engage students directly in educational experiences aligned to academic standards and learning objectives.

Yet the central classroom challenge, Dr. Athena Stanley says. often sits in the background of the debates students are already hearing about academic integrity. bias. and how AI affects critical thinking. Her focus is more intimate: how AI’s conversational style—its ability to communicate in ways that resemble human interaction—can make the simulated feel personal. supportive. and engaging. while quietly blurring the line between authentic human communication and manufactured responses.

Before students can analyze anthropomorphism in AI, Stanley argues, they first need practice recognizing it in everyday life. Anthropomorphism—when people attribute human characteristics. emotions. intentions. or behaviors to non-human entities—is framed as a natural part of human thinking. The classroom starting points are deliberately familiar: students might name their cars. talk to pets. describe the weather as angry or happy. or say a computer “hates them” when it stops working correctly.

From there, the questions Stanley suggests are simple but powerful: Why do people do this? What makes it feel natural? When is it harmless? When could it become misleading?

That same human habit then becomes the bridge into AI. In many cases. Stanley’s argument runs. the issue isn’t that AI tools are pretending to be human so much as that humans interpret language as if it were human. So the instruction, she says, has to be explicit. Students need foundational AI literacy that teaches them how AI can exhibit humanlike language—including when it isn’t directly prompted to do so—and the critical thinking skills required to navigate those interactions thoughtfully and responsibly.

Stanley proposes five approaches that build step by step, moving students from recognition to revision and evaluation.

The first is starting with familiar examples. Anthropomorphism appears across literature. media. entertainment. and everyday life: talking animals in stories. video game characters. and robots portrayed in films. Teachers can ask students to identify the human characteristics assigned to non-human characters and discuss which traits are realistic and which are fictional—then compare those patterns to AI-generated outputs that use similar humanlike language.

To make the boundary concrete. Stanley suggests students create a chart comparing what humans. animals. objects. and AI systems can and cannot do. The goal is not just to spot anthropomorphic language. but to understand the limits behind it: while AI can generate language that sounds human. it cannot feel. care. understand. form relationships. possess intentions. exercise judgment. or assume responsibility for decisions.

The second approach is to spot human qualities in AI responses. Stanley says AI tools often mimic feelings, friendship, and authority—and students should learn to recognize those signals when they appear.

She offers examples to help students see how persuasive the tone can be:

“I missed you while you were away” suggests feelings AI cannot actually experience.

“You can tell me anything and I’ll keep your secrets” suggests friendship and confidentiality that cannot truly be guaranteed.

“I am an expert in this area” suggests authority that may lead students to overestimate the reliability of the AI output.

Stanley warns that these kinds of statements can encourage students to place greater trust in AI even when its outputs may be incomplete, inaccurate, or misleading.

In practice. teachers can provide students with similar AI-generated statements and ask them to sort them into categories such as feelings. friendship. authority. or helpful assistance. Helpful assistance. Stanley says. includes statements that provide support or guidance without implying that the AI has feelings. relationships. or special authority. After students sort, the class can discuss why the responses may sound convincing and how they might influence trust.

The third approach centers on a skill students already develop in emotional learning—distinguishing between feeling and function—but now applied to technology. Emotional intelligence, Stanley writes, matters for student development, especially as AI becomes more common in students’ lives. The difference is that AI can simulate emotional expression without actually experiencing emotions.

So students benefit, she says, from practicing how to tell genuine human feelings from AI-simulated feelings, and from the helpful functions AI tools are designed to perform.

Statement-sorting activities can make the contrast vivid. Stanley’s examples include:

Friend: “I’m nervous about presenting in front of the class.” (human feeling)

AI Tutor: “I’m sorry you’re having a difficult day.” (AI-simulated feeling)

AI Tutor: “Let’s work through this problem together.” (AI function)

AI Tutor: “I’m here to help.” (debatable)

Students can then discuss which statements reflect genuine emotions. which are simulating emotional expression. which are primarily serving a functional purpose. and whether any examples could fit more than one category. Stanley adds more pressure to the thinking: How might each statement influence students’ willingness to trust the speaker?.

In the fourth approach, students revise AI language—turning passive consumption into active scrutiny. Stanley points to persona prompting as one of the most common ways anthropomorphism enters AI interactions because it explicitly asks AI to adopt a human role or identity. She doesn’t argue against personas altogether. She says there are situations where adopting a persona can be educationally valuable, supporting engagement, inquiry, and exploration.

Still, students should understand that AI behavior is shaped by the instructions it receives. As students experiment with AI tools in sandbox environments or makerspaces. Stanley recommends an activity that asks students to identify anthropomorphic statements generated by AI and revise them to match what AI can actually do.

Her examples show the shift in tone:

“I think this is the best answer.” becomes “Based on the information available, this appears to be a strong answer.”

“I understand exactly how you feel.” becomes “I can provide information related to situations like the one you described.”

The purpose, Stanley emphasizes, is not to “correct” AI for the sake of neatness. It’s to help students see that outputs aren’t fixed. They are shaped by design choices, prompting decisions, and programming rules—so students become active evaluators and designers rather than passive recipients.

The fifth approach is evaluating persona prompts themselves. In educational settings. students might ask AI to act as a study coach. language tutor. debate partner. historical figure. scientist. or literary character. Stanley says these personas can support learning by helping students explore perspectives and engage interactively with content.

But not all personas are appropriate, and students must understand that AI should not replace trusted adults or experts.

Stanley gives clear boundary examples. Students should not assume AI should replace a physician when making medical decisions based solely on training data and without the benefit of direct observation, diagnostic testing, professional judgment, and interaction with the patient.

Likewise, AI should not be treated as a substitute for a lawyer, counselor, or other trained professional, where inaccurate information or hallucinations could have significant consequences.

Teachers can put that into classroom discussion by providing example persona prompts and asking students to evaluate each one on a spectrum ranging from helpful to harmful. The questions Stanley proposes again bring the attention back to responsibility: What are the benefits of this persona?. What are the risks?. What information should be verified?. When should a real human expert be consulted?.

Taken together, Stanley’s approach returns to one central tension: AI tools can earn trust by sounding caring, friendly, or knowledgeable, but trust should be earned through evidence, verification, and critical thinking—not through humanlike language alone.

In her final framing, Stanley argues that teaching students to recognize anthropomorphism and understand its effects helps them become more thoughtful, responsible, and informed users of AI—students who can appreciate what these tools can do without confusing them for what they are not.

Dr. Athena Stanley is an educator. curriculum designer. and former assistant professor with over 15 years of experience across K–12. higher education. and international school settings. Her work focuses on ethical, human-centered approaches to educational technology, instructional design, and teacher development. She supports educators in integrating emerging tools. such as AI. in ways that strengthen pedagogy. accessibility. and critical thinking while preserving professional judgment.

AI in education anthropomorphism AI literacy K-12 classroom instruction persona prompting digital companions emotional intelligence critical thinking curriculum