Chat with George Washington: A Classroom Activity Guide

In a 2025 internal study, Humy analyzed 3,677 student conversations with the AI George Washington across 34,067 messages to surface what students actually asked when given a source-grounded conversational partner. The patterns were specific: the largest share of conversations (1,200, or roughly a third) covered Washington’s life, presidency, and accomplishments; another 900 turned on his Revolutionary War military leadership; 800 engaged the Constitution and the Convention; 700 went into his personal life and family; and 600 explicitly took on slavery and his views of it. Grade 8 produced the heaviest use, with 139 teachers leading conversations in that band alone. The data answered a question I had not been able to answer cleanly before, which is whether students given a chat with Washington reach for substantive history or for trivia. They reach for substantive history almost all of the time, and the proportion that engages slavery on Washington’s own record is the finding that should reshape how the unit is taught.
This is a classroom activity guide for a Washington chat across three grade bands (5th, 8th, and 11th), built around the documentary record curated by George Washington’s Mount Vernon and the Gilder Lehrman Institute . The structure of each activity is the same; what changes is the document set, the chat depth, and the writing output the activity produces.
Why Washington specifically
Two reasons. The first is that the documentary record is unusually deep. Washington’s correspondence runs to roughly 140,000 documents preserved across Mount Vernon’s archive, the Library of Congress, and the Papers of George Washington at the University of Virginia. A source-grounded AI Washington has enough underlying material to support careful sourcing and contextualization work, which is what makes the chat pedagogically useful rather than a novelty.
The second is that Washington’s presidency is anchored in decisions that map cleanly onto the C3 Inquiry Arc. Decisions about the Whiskey Rebellion, the Jay Treaty, the Proclamation of Neutrality, and the Farewell Address are concrete cases where students can practice asking why this and not that, which is the historiographical question the discipline cares about. They are also the questions students actually ask, per our conversation data.
The 5th-grade activity: Washington as a person
A 5th-grade unit on the American founding usually has a narrow window for primary-source work, so the activity is short and the document set is curated tightly.
Open with a single primary-source artifact: a page from Washington’s 1789 inaugural address , reproduced at a 5th-grade reading level. Students read it aloud as a class, with the teacher pausing on two or three vocabulary words.
Open a Humy chat with the AI George Washington and have students take turns asking one prepared question each. Useful starter prompts: “What was the hardest day of the Revolutionary War for you?” “Why did you decide to step down after two terms?” “What did Mount Vernon look like when you were home?” The teacher keeps the chat constrained to roughly 10 minutes, with the figure responding from the documentary record Mount Vernon publishes.
After the chat, students complete a one-paragraph writing prompt: pick one thing Washington said that surprised you, and explain why. The chat is the rehearsal; the writing is the assessment.
The 8th-grade activity: the Whiskey Rebellion as a decision
8th grade US history typically covers the early republic with more analytical depth, and the Whiskey Rebellion is the canonical decision point for a Washington chat.
Hand students a packet of three primary sources: an excerpt from Washington’s 1794 Proclamation against the Whiskey Rebellion , a contemporaneous farmer’s pamphlet, and a short modern secondary source on the political stakes. Students read and annotate before any AI use.
The chat workflow runs the four-phase pattern we cover separately. Students open a Washington chat and ask HIPP-style sourcing questions about the Proclamation: who was Washington addressing, what was happening politically, why did he choose to lead troops in person rather than send a deputy, what was the alternative he rejected. The teacher coaches students to push past the first answer into the why this and not that layer.
A short corroboration step: students open a second chat with Alexander Hamilton, who has his own AI figure in the platform, and look for where Hamilton’s framing of federal authority differs from Washington’s restraint. The contrast is the lesson.
Students write a one-page response: was Washington’s decision to lead troops the right call given what he knew at the time? Defend with evidence from the documents and the chats.
The 11th-grade / AP US History activity: Washington and the precedent problem
AP US History expects students to do real historiographical work, and Washington is one of the cleanest case studies in the course for the question of how does a precedent become a tradition.
The activity centers on Washington’s Farewell Address and its position in APUSH Period 4 (1800–1848) as well as the foreign-policy through-line into Period 6 (1865–1898) and Period 7 (1890–1945). Students get the address itself, an excerpt from the Proclamation of Neutrality , and a more recent scholarly excerpt on the long influence of Washington’s foreign-policy posture.
Students open a Washington chat focused on the Farewell Address. The questions they practice are higher-order: why did Washington address party rivalry as the foundational threat? What did he see in the Jay Treaty debate that pushed him toward the warning he wrote? How did he understand the relationship between domestic faction and foreign entanglement?
A corroboration step with the AI Thomas Jefferson, on his own foreign-policy positions from the same period, sharpens the historiographical move.
The writing output is a DBQ-style 5-paragraph essay, with the chat transcripts available to students during the writing as a reference for their sourcing analysis, but with the actual argument formation belonging to the student.
Common patterns across the three activities
Three patterns hold across grade bands and across the 3,677-conversation dataset.
The first pattern is that students lean toward substantive history rather than trivia when the activity is framed correctly. The conversation data shows that the legends most teachers worry students will ask about (the cherry tree, the wooden teeth) ended up grouped into a single “myths and legends” category of 600 conversations out of 6,675 thematic appearances. That is meaningful but a small fraction next to the presidency category at 1,200 and the Revolutionary War category at 900. When students get a real conversational partner, they reach for real questions, and the conversations themselves do not look like the textbook treatment.
The second pattern is that the chat surfaces student misconceptions earlier than reading alone does. Washington as a person rather than a marble statue is the most common misconception worth surfacing, and 10 minutes of conversation with the figure does what a unit of reading often does not. The biggest single-grade adoption in our dataset is grade 8 (139 teachers), which is exactly the grade where the marble-statue framing is hardest to dislodge.
The third pattern is that slavery shows up in student conversations whether the teacher invites it or not. The dataset’s 600 conversations explicitly on slavery and Washington’s slaveholding (more than on his personal life, more than on his interactions with other Founders) suggest that the curricular emphasis the guardrails section below names is matched by what students themselves want to engage with. The Mount Vernon primary source education collections , including the 1793 spinning report that names ten enslaved women at Mount Vernon , and the Be Washington decision simulation pair naturally with an AI Washington chat that is honest about Washington’s record.
Guardrails specific to Washington
A Washington chat touches two histories that need explicit framing: slavery at Mount Vernon, and the Indigenous nations whose lands Washington’s policies affected.
Slavery is the first, and it sits at the center of the unit rather than at its margin. Washington enslaved 317 people at Mount Vernon by the end of his life, and Mount Vernon’s own teaching resources on slavery treat the enslaved community as named historical actors with documented work, family, and resistance histories rather than as background to Washington’s biography. The platform’s AI Washington should be configured to engage that record honestly, including his complicated final-decade movement on the question (the will-based manumission of the 123 people he personally owned, which took effect only after Martha’s death) and the limits of that move (the dower slaves of the Custis estate remained enslaved). The data already shows students asking about it. The job of the platform is to give them a Washington who can answer in his own documented words rather than in a sanitized one.
Indigenous history is the second framing imperative. Washington’s administration negotiated treaties (Treaty of New York with the Creek, Treaty of Holston with the Cherokee, Treaty of Greenville with the Northwestern Confederacy after Fallen Timbers) on terms that set the precedent for nineteenth-century federal Indian policy, and the K-12 record of how those negotiations are taught has not kept up with current scholarship. The Smithsonian’s Native Knowledge 360 program is the strongest current grade-banded curriculum for these histories, with explicit tribal-voice priorities, and any AI Washington unit that touches Indigenous policy should pair the chat with NK360 lesson materials rather than running the chat alone.
Putting the activity in your unit
If you teach US history and want to drop one of these activities into the unit you are running this month, try Humy free and run the grade-appropriate version with one section first. The fastest way to evaluate the activity is to see what students actually ask, not what a sample transcript suggests they might.