Artificial intelligence is changing how students encounter history. Instead of simply reading about Winston Churchill, Rosa Parks, or the Industrial Revolution, students can now explore the past by asking questions to AI-generated representations of historical figures.
Imagine a student interviewing a factory worker in nineteenth-century Britain, questioning a civil rights activist in 1960s America, or discussing a difficult wartime decision with a simulated political leader. These conversations can make history feel more immediate and encourage students to explore perspectives they might otherwise overlook.
But they also raise an important educational question: How can we ensure that students learn to investigate history rather than simply believe what an AI character tells them?
AI historical role-play has educational potential, but only when it is designed around evidence, critical thinking, and teacher guidance. The goal should not be to recreate the past perfectly. It should be to help students develop the skills needed to question historical claims and construct their own interpretations.
Platforms such as Parlova.org are exploring this approach by allowing students to investigate historical questions through conversations with AI-generated characters, while giving teachers the tools to guide and monitor the learning process.
Here are five practical approaches teachers can use.
1. Start With a Historical Mystery, Not a Conversation
One common mistake when introducing AI historical characters is to let students chat without a clear objective.
Although an open conversation can spark curiosity, it does not necessarily lead to meaningful historical understanding.
A better approach is to begin with an enquiry question.
For example:
- Why did some American colonists support independence while others remained loyal to Britain?
- What motivated individuals to participate in the American civil rights movement?
- How did the Industrial Revolution change the lives of factory workers and their families?
Students can then interview simulated historical figures to gather possible explanations.
For a lesson about the American Revolution, one group might question a Patriot merchant, while another interviews a Loyalist or a British official.
Their task is not simply to collect answers. It is to understand why people living through the same events could interpret them so differently.
This transforms the activity from casual conversation into a structured investigation. Students have a problem to solve, competing perspectives to explore, and conclusions to justify.
The teacher’s role is to establish the historical question and explain what constitutes a convincing, evidence-based answer.
2. Treat Every AI Answer as a Claim to Investigate
An AI-generated historical character may sound persuasive, confident, and historically informed. But that does not make its statements reliable.
Language models can produce factual errors, invent details, or present contested interpretations as established facts.
Rather than hiding this limitation, teachers can turn it into an educational opportunity.
After receiving an answer, students can ask themselves:
- What factual claims did the character make?
- Which claims can be verified?
- What evidence would support or contradict them?
- Is the character expressing a particular perspective or describing an established fact?
For example, a simulated Winston Churchill might offer a convincing explanation of a controversial wartime decision. Students should then compare that explanation with contemporary speeches, government records, and historians’ interpretations.
The comparison might reveal that the character has simplified a complex situation, omitted important context, or attributed motives that historians continue to debate.
This distinction is essential: An AI character is a learning simulation, not a primary historical source.
Even when its answers are accurate, they are generated from a modern computational model rather than recovered directly from the past.
The most valuable part of the exercise may therefore be the moment when students discover that a plausible answer is not necessarily a correct one.
3. Introduce Contradictory Perspectives
History rarely provides a single, uncontested interpretation of events.
AI role-play can help students explore this complexity by presenting different historical viewpoints.
Consider a lesson about the Industrial Revolution.
Students might question a factory owner, a textile worker, and a social reformer, each responding from a different economic or social perspective.
A factory owner might emphasise economic growth and technological progress, while a worker might describe long working hours, unsafe conditions, and the challenges of urban life.
A reformer might focus on child labour, education, and the need for new legislation.
Students can compare these arguments and identify the interests, assumptions, and historical circumstances behind them.
The same approach could be used to investigate the American civil rights movement, European colonial expansion, or the social consequences of industrialisation in different parts of the world.
However, teachers should distinguish between historically documented viewpoints and dialogue invented by the AI to illustrate them.
The objective is not to pretend that we can recover exactly what historical people would have said. It is to explore perspectives grounded in historical evidence while recognising the limitations of simulation.
This approach also helps students understand that disagreement does not mean all interpretations are equally well supported.
4. Give Teachers Visibility Into the Investigation
AI tools are sometimes presented as independent tutors capable of replacing direct teacher involvement.
For historical enquiry, that would miss an important part of the educational process.
Teachers need to understand not only what conclusions students reach, but also how they reach them.
Did a student ask relevant questions? Did they challenge a suspicious answer? Did they distinguish between evidence and interpretation? Were they misled by an invented historical detail?
These observations can reveal misconceptions that might otherwise remain invisible.
This is one of the principles behind Parlova, a classroom platform that lets students investigate historical questions through conversations with AI-generated characters.
Teachers can create guided missions, follow students’ conversations in real time, and review their responses afterwards. The platform also provides information about student progress and answers, helping teachers identify which ideas have been understood and which may require further discussion.
For example, if several students accept the same inaccurate claim about the causes of the First World War, the teacher can revisit that point collectively and examine the historical evidence.
Instead of simply marking answers as correct or incorrect, the teacher can use the activity to understand how students reasoned and where their interpretations diverged.
The technology becomes a way to make students’ thinking more visible, rather than a substitute for teaching.
5. Finish With an Evidence-Based Debrief
A successful historical role-play should not end when the conversation ends.
The final stage is where students transform their exploration into historical understanding.
One simple activity is to ask each student or group to produce a short conclusion answering the original enquiry question.
Their answer should include three elements:
- A clear historical claim.
- Evidence supporting that claim.
- An explanation of any uncertainty or competing interpretation.
For example, after investigating working conditions during the Industrial Revolution, students might be asked:
“Did industrialisation improve the lives of ordinary people?”
Different groups may reach different conclusions depending on the period, region, and social group they investigated.
One group might emphasise rising production and long-term improvements in living standards. Another might focus on dangerous working conditions, child labour, and overcrowded cities.
The teacher can then ask students to defend their interpretations using historical evidence rather than relying on what their AI characters said.
Which students consulted additional sources? Which relied too heavily on the simulation? Did different groups reach different conclusions, and why?
A classroom discussion can highlight both sound historical reasoning and the limitations of generative AI.
This also creates an opportunity to teach broader digital literacy skills that extend beyond history: evaluating sources, questioning confident assertions, and recognising uncertainty.
From Historical Simulation to Historical Thinking
AI historical role-play is not valuable simply because students can talk to a simulated historical figure.
Its educational potential comes from what students do with that interaction.
When teachers provide a clear enquiry question, encourage source verification, introduce competing perspectives, and organise a structured debrief, AI dialogue can become a starting point for meaningful historical investigation.
This approach still requires careful evaluation. Engaging activities do not automatically produce better learning outcomes, and the effectiveness of AI-supported historical enquiry needs to be tested in real classrooms.
Nevertheless, the underlying educational principle is straightforward: students should learn to question the information they receive, including information produced by artificial intelligence.
Perhaps the most useful question we can teach students to ask an AI historical character is not simply, “What happened?”
It is: “How do we know?”
Quentin Kral is an astrophysicist at the Paris Observatory and the founder of Parlova, the platform mentioned in this article.
