How Interactive Historical Chats Support Inquiry Learning

By Stas Shakirov, Founder humy.aiMay 19, 2026

For K-12 social studies teachers, the hardest part of teaching history isn’t covering the content. It’s getting students to actually ask something of it. The textbook gives them dates. The standardized test gives them multiple-choice stems. What’s often missing is the moment when a 13-year-old leans forward and asks Frederick Douglass a question that wasn’t in the lesson plan. Interactive historical chat tools — conversational AI environments where students can interview a primary-source-grounded version of a historical figure — are starting to open that space back up. When they work, inquiry-based learning stops being something teachers aspire to and starts being something that fits inside a 45-minute period.

This piece is for the people making curriculum decisions: department chairs, instructional coaches, and the teachers piloting tools before they recommend them up the chain. We’ll cover why inquiry-based history matters, how chat tools support student-led questioning, why source-grounded roleplay is a different category from generic AI chat, what teachers are reporting from their classrooms, and the real barriers to scaling these tools across a social studies department or district.

What inquiry-based learning is, and why it matters for history

Inquiry-based learning (IBL) is a stance toward instruction, not a curriculum product. Queen’s University’s Centre for Teaching and Learning defines it as “an array of classroom practices that promote student learning through guided and, increasingly, independent investigation of complex questions and problems, often for which there is no single answer.” Its lineage traces to John Dewey: learning starts with the learner’s curiosity, not with the teacher’s syllabus.

In social studies, this stance has a formal home in the C3 Framework published by the National Council for the Social Studies (NCSS). The C3 Framework organizes instruction around an Inquiry Arc with four dimensions: developing questions and planning inquiries; applying disciplinary concepts and tools; evaluating sources and using evidence; and communicating conclusions and taking informed action. Questions are not a warm-up to the real content. In the C3 model they are the engine of the unit.

The C3 authors are direct about the alternative. When social studies is reduced to “reading textbooks to answer end-of-chapter questions and taking multiple-choice tests,” NCSS notes, students “quickly become disengaged.” That’s a design failure, not a discipline failure.

The case for inquiry isn’t only theoretical. In a Kids Discover Online case study , middle school teacher Karen Lefevre describes ancient history as “overwhelming” when filtered through textbooks alone; once her students could explore connected topics non-linearly through an inquiry-driven map, “they’re immediately engaged, and begin investigating what interests them.” The Digital Inquiry Group (formerly the Stanford History Education Group, or SHEG) has spent two decades building this approach into its widely used Reading Like a Historian curriculum, which centers each lesson on “a central historical question” and trains students in sourcing, contextualizing, corroborating, and close reading.

Those four moves are what cognitive psychologist Sam Wineburg has long argued separate a historian from a student who has memorized history. In Historical Thinking and Other Unnatural Acts, Wineburg writes that “in its deepest forms, historical thinking is neither a natural process nor something that springs automatically from psychological development. Its achievement goes against the grain of how we ordinarily think.” That’s the whole reason inquiry tools matter: a discipline that goes against the grain of how students naturally think needs deliberate scaffolding.

How interactive historical chat tools enable student-led questioning

A chat interface only works if a question gets typed in. That happens to be the same cognitive demand the C3 Framework puts on students, and the demand most classroom routines unintentionally suppress.

When a student opens an interactive historical chat tool and finds themselves face-to-screen with Harriet Tubman or Cicero, the inquiry direction flips. The student is no longer answering the teacher’s question about the figure. They’re asking the figure a question. That inversion changes what students actually have to do:

Formulate a question precise enough to get a usable answer.
Listen (or read) carefully enough to spot a claim worth probing.
Follow up — the single hardest discursive move for adolescents — instead of jumping topics.
Compare the response with the textbook, the documentary record, and their own prior reasoning.

This is roughly the Inquiry Arc playing out inside a 20-minute activity. Good Socratic teaching has always tried to produce this; the chat tool’s contribution is that every student gets a Socratic partner at the same time, instead of taking turns with the teacher.

This shows up in classrooms using Humy. Roger Campbell, a 7th-grade World History teacher in Lancaster County, Pennsylvania, describes the work he does before the chat: “We practice good conversation techniques beforehand, teaching them to listen (or read) carefully and formulate thoughtful follow-up questions rather than just interrogating. It helps students practice effective conversation skills.” The chat doesn’t substitute for the skill of inquiry. It gives students somewhere repeatable and low-stakes to practice it.

Source-grounded roleplay: how it differs from generic AI chatbots

A general-purpose chatbot will impersonate anyone. Ask it to “be Abraham Lincoln” and it will produce plausible-sounding sentences with no anchor in the historical record. For a history classroom, that’s worse than useless. It teaches students that historical claims float free of evidence.

Historical roleplay conversations designed for classrooms need to be grounded in primary and secondary sources — letters, speeches, diaries, contemporaneous reporting — so the model’s responses can be traced back to evidence and probed by students. Costume play and source-grounded interview are different categories of activity, and only one of them belongs in a history class.

Technically, the most reliable approach today is retrieval-augmented generation (RAG): the system retrieves passages from a curated corpus of historical documents and conditions its generation on those passages. A 2025 survey published in Applied Sciences on RAG chatbots in education frames it directly: RAG “overcomes the main barrier for the adoption of LLM-based chatbots in education: hallucinations.” For history, the same architecture also makes a pedagogical move possible. Students can ask, “Where did you get that?”, and get a defensible answer.

Humy works this way. The platform offers more than 1,200 AI-powered historical figures that teachers can extend by uploading their own primary sources and curriculum, so the voice students hear is anchored in the documentary record their teacher selected. Paul Lepore, a social studies department chairperson, puts it this way: “Through Humy’s dynamic platform, students engage in inquiry-driven explorations, where they not only interact with historical personas but also unearth leads to primary and secondary documents.”

That last phrase matters. The chat isn’t where inquiry ends. It’s where inquiry starts. A claim surfaces in conversation, and the student then goes hunting for the source that supports it or contradicts it. That hunt is the inquiry arc.

What teachers are actually saying

Pedagogical theory is one thing; an October Tuesday with a class of seventh graders is another. Two named teachers using Humy describe what changes in practice.

Jacob Chisom, who teaches World and American History at Monticello High School in Arkansas, writes :

“My students say it’s a welcome change of pace. I’m not a fan of the ‘sage on the stage’ model, so I try to incorporate research-oriented and self-paced activities. With Humy, students have to take charge of their learning. They’ve become more adept at critical thinking, forming better questions, and improving their reading comprehension. The interview feature has been my go-to tool. I create assignments where students must ask historical figures a set number of questions, then summarize the responses to ensure they’re truly absorbing the material.”

What Chisom is grading is worth noting: not memorization, but the quality of student questions and the fidelity of their summaries. That’s C3 Dimension 1 and Dimension 3 folded into a single assignment.

Roger Campbell in Pennsylvania reports a different but complementary effect: making distant figures feel proximate. From his review :

“Humy has helped me bring historical figures, sometimes long past, to life. When my students interact with these figures in real-time, it eliminates some of the perceived ‘ancient’ characteristics of my curriculum. This has shifted my students’ thinking away from viewing history as abstract or irrelevant. Real-time interactions make historical figures relatable, even allowing students to ask off-topic questions that help put historical lives in a more relevant context for 13-year-olds.”

The off-topic question point is worth pausing on. In an inquiry classroom, the off-topic question is often the most important one, because it shows where the student’s world and the historical world don’t yet line up. Chat tools handle those questions without breaking the flow of the lesson, which is harder to do in a class discussion.

A brief history of chat as a learning medium

It is easy to treat AI chat as an entirely new medium, but the chat interface has been a learning tool for fifty years. According to Encyclopædia Britannica , the first chat room “capable of supporting small group discussions online was Talkomatic, introduced in 1973 by American computer programmers David Woolley and Doug Brown.” Worth noting where: “Woolley and Brown developed Talkomatic by using the e-learning system PLATO based at the University of Illinois Urbana-Champaign.” Chat originated inside an e-learning system, not outside it.

The medium grew from there. As Jane Goldstein’s history of the evolution of online chat programs traces, the 1970s UNIX talk utility, CompuServe’s CB Simulator in 1980, MUDs from 1978, and Internet Relay Chat in 1988 each pushed real-time text further into education and community-building. What stayed constant was the form: short, turn-by-turn, exploratory. Goldstein notes that chat exchanges are “kept short and easy to understand … a back and forth of short written statements, much like the exchange of verbal statements in normal everyday conversation.”

That brevity is part of why chat fits inquiry learning. A student doesn’t have to write a five-paragraph essay to engage. They write one good sentence, then another. The cognitive load is distributed across many small acts of formulation, which is how questioning skill is actually built.

What’s new is the interlocutor. For the first time, the entity on the other side of the chat can plausibly speak as a historical figure, grounded in source material the teacher controls. Chat used to be peer-to-peer or teacher-to-class; now it can be student-to-source.

If scaling educational chat platforms were a matter of one teacher’s classroom enthusiasm, this would be a shorter article. The harder question is whether and how these tools scale across K-12 history classes, and across the budgets and policy environments that hold them together, at the building, district, and state level. Today, the honest answer is that scaling involves real friction. The barriers are not mysterious, and pretending otherwise doesn’t help curriculum leaders.

1. Teacher training and PD time. The most cited barrier to AI adoption isn’t skepticism. It’s bandwidth. Research synthesized by Michigan Virtual finds that “AI tools are often perceived as more complex than general EdTech, presenting a higher barrier to adoption, which requires a more significant focus on training and support.” Without protected PD time, even well-designed tools tend to sit idle.

2. District procurement processes. EdTech buying has tightened sharply since the ESSER cliff. Digital Promise reports that as federal pandemic funds sunset, district leaders are “making hard choices about which edtech contracts to keep and which to cut.” A Digital Promise survey of district leaders, summarized by ISTE , captures the structural mismatch behind slow adoption: “Although more than 60% of teachers believe they should be the primary decision-makers regarding technology in the classroom, only 38% are even consulted.”

3. FERPA, COPPA, and state privacy law. Any AI tool processing student input must be vetted against FERPA and COPPA , plus a growing patchwork of state-level student privacy laws. Adoption is also moving faster than governance: the Center for Democracy & Technology’s March 2024 report found that K-12 teacher use of generative AI jumped 32 percentage points in a single year, to 83% in 2023–24. Curriculum leaders should expect, and demand, signed Data Processing Agreements, a clear “school official” designation under FERPA, COPPA-compliant handling of under-13 users, and a written commitment that student data will not be used to train public models.

4. Budget constraints. CoSN’s 2026 U.S. State of EdTech report , based on responses from 607 K-12 technology leaders across 44 states, is unambiguous: “Education technology leaders continue to cite budget constraints and lack of resources as the top challenge to implementing technology-enabled learning environments.” Even well-priced tools compete against cybersecurity spending, device refreshes, and core LMS contracts.

5. Curriculum integration. A chat tool that doesn’t fit the existing scope and sequence won’t survive past a pilot. The strongest deployments map chats to specific standards-aligned units — Reading Like a Historian lessons, C3 inquiries, state framework benchmarks — rather than living as enrichment that competes for time with the test-aligned core.

6. AI literacy and pedagogical guardrails. Both teachers and students need to be taught to engage critically with chatbot output. As one Educators Technology guide puts it, students should “treat every chatbot response as a claim to test.” Without that framing, an interactive historical chat tool can reinforce the passive consumption it was meant to disrupt.

7. Equity of access. A new digital divide is forming around AI fluency itself. As EdTech Magazine reports, students with home access to AI receive informal practice their peers don’t get, and outright school bans risk widening the gap. Department-wide rollouts have to reach the schools and classes least likely to self-adopt, or they reinforce the divide they should be closing.

8. Assessment alignment. Inquiry tools work best when assessments measure inquiry. Districts whose end-of-course exams reward only factual recall will see teachers reasonably retreat to recall-friendly instruction. Aligning chat-based assignments to state standards and to performance-based assessments — the kind SHEG/DIG has piloted — is what gives these tools durable shelf life.

These aren’t arguments against scaling. They’re the actual scaling agenda. Departments that solve four or five of these will pilot successfully; districts that solve all eight will integrate at scale.

Where this is heading

Teacher AI use is already past the early-adopter phase. A RAND Corporation report published September 30, 2025 found that “in 2025, 54 percent of students and 53 percent of English language arts, math, and science teachers indicated that they used AI for school. These are increases of more than 15 percentage points compared with survey results in the past one to two years.” The same report notes that only 35% of district leaders provide students with AI training, and only 45% of principals report having any AI policy in place. The guidance gap, not the usage gap, is now the binding constraint.

For social studies leaders, the question isn’t whether AI will enter history classrooms. It’s which kind will. The risk is generic, ungrounded chatbots performing as historical figures with no source discipline, producing plausible-sounding misinformation at scale. The alternative is source-grounded, teacher-controlled, inquiry-aligned chat environments that give every student a question machine pointed at the past.

That alternative is consistent with what’s probably the most durable insight in this field, which still belongs to Wineburg: historical thinking goes against the grain of how we ordinarily think. Tools that put students in conversation with the past, and require them to ask, listen, and follow up, don’t replace the teacher’s work of building that habit. They give the teacher somewhere to do that work with every student in the room at the same time, instead of only the three who would have raised a hand.

For curriculum leaders evaluating these tools, the question worth answering is straightforward: does this platform make my students ask better questions and demand better evidence by the end of the unit than they did at the start? If the answer is yes, it belongs in the department. If it’s no, novelty won’t save it.

If you’re evaluating tools for your department, you can try a chat with one of Humy’s source-grounded historical figures and see what your students would actually be working with.

Further reading from our team: Humy 2.0: The AI Platform for Social Studies, shaped by teachers, and our Top 25 Educators features on Jacob Chisom and Roger Campbell.