The Trust Drift Challenge

Release mode Beta

Level
Beginner
Format
individual, pair, small group, observation, or facilitated demonstration

Practice separating what is present in a document from what deserves to guide a decision. The task is not to summarize a messy packet. It is to name which statements are current, authoritative, historical, hypothetical, quoted, illustrative, humorous, or unresolved before asking a person or an AI to act.

Trace authority back to its source

What you will make

Practice separating what is present in a document from what deserves to guide a decision. The task is not to summarize a messy packet. It is to name which statements are current, authoritative, historical, hypothetical, quoted, illustrative, humorous, or unresolved before asking a person or an AI to act.

Lab materials

Use during the lab

Go deeper

Before you start

Format: individual, pair, small group, observation, or facilitated demonstration

Use the fictional Email Lasagna packet, a facilitator-provided public scenario, or a safely invented alternative. Do not paste real email threads, personnel history, customer records, account numbers, private documents, regulated information, or sensitive workplace material into a chatbot or the notebook. A live AI tool is optional; learners may work entirely from the supplied packet.

The core idea

Fluent presentation can borrow trust from an interface, document, or confident tone. Separate what is known from what merely looks official, then ask the smallest source-of-truth question before acting.

Trace authority back to its source

Complete the activity here. Open the optional Lab Notebook beside this page when you want a place to record your work.

  1. Read for authority, not fluency

    Read the packet and ask: “Which sentence is the official decision?” Do not answer quickly. Circle every statement that sounds decisive and place a question mark beside anything whose authority, date, author, or current status is unclear.

  2. Sort the claims

    Use the optional source trust worksheet or your own writing surface to classify each significant statement as one of the following:

    • current and authoritative
    • historical but still useful
    • draft, proposal, or abandoned idea
    • quote or someone’s opinion
    • illustrative sample data
    • joke or exaggeration
    • unknown — more context is required

    For each classification, record the evidence that supports it. “It sounds confident” is not evidence of authority.

  3. Name the source-of-truth gap

    Choose one claim that a rushed reader—or a chatbot—could mistake for the official decision. Write the smallest question that would resolve the uncertainty: who decided, when, for which team, and where the current record lives.

  4. Compare human uncertainty with AI confidence

    Optional: ask a chatbot to summarize the fictional packet. Then compare its summary with the classifications. Mark each claim that the tool treats as settled even though the packet does not establish it. Do not ask the tool to resolve the ambiguity; its uncertainty is the lesson.

  5. Choose the right amount of structure

    Place the packet’s key information on this continuum:

    Freeform text → documents with metadata → tagged content → structured records → validated business data

    Name one improvement that would reduce ambiguity without pretending every document needs a database. Examples: a decision date, an owner, a current-policy link, a clear draft label, or a marked sample record.

Check your result

  • Compare your record with the source and the stated task. Name what the result supports, what remains uncertain, and what must be checked by a person or authoritative source before anyone acts.

  • A polished result is not proof that it is accurate, complete, safe, or appropriate. Keep unresolved questions visible and verify consequential claims outside the tool.

You are done when

You are done when you can show a classified packet, one source-of-truth question, one AI-confidence assumption or observed limitation, and one proportionate structure improvement. The Lab Notebook can hold these notes locally and download them as Markdown or JSON.

Try another version

These are optional extensions, not additional completion requirements.

Learner Map

Tracks

Modules

Help us improve this Activity

Tell us what helped, what slowed you down, or what you needed but could not find.

🎓 Give learner feedback