L8 EXERCISE: ELICIT, VERIFY, CORRECT — SUBMISSION
Label: AI allowed — keep a record

Topic number and name:
Date(s):

======================================================================
PART 1. EXPECTATIONS (write this BEFORE you send the prompt)
- What do I already believe about this topic?
- Which kinds of claims do I expect to be hardest to check?


======================================================================
PART 2. THE ELICITED ANSWER
Paste the assistant's full answer exactly as received. Do not edit it.
If you asked follow-up questions, paste those exchanges too.


======================================================================
PART 3. VERIFICATION LOG
One row per checkable claim. EVERY checkable claim, not just the ones
that look wrong or are easy to check. Number them.

Result words: CONFIRMED / WRONG / NOT FOUND / COULD NOT CHECK

 #  | Claim (quote it)        | Authority (source)   | Where I looked        | Result  | What the source says
----+-------------------------+----------------------+-----------------------+---------+---------------------
 1  |                         |                      |                       |         |
 2  |                         |                      |                       |         |
 3  |                         |                      |                       |         |
(add rows)

Claims I could not check, and why:


======================================================================
PART 4. ERROR RECORD
For each WRONG or NOT FOUND claim:
 - Claim #:
 - What was wrong:
 - How I detected it (which source, which line of it):
 - Which failure mode it looks like (hallucination / sycophancy /
   miscalibration / staleness / other) and the mechanical cause:


======================================================================
PART 5. CORRECTED PRODUCT
The final version you would be willing to sign for. Every claim in it
must be CONFIRMED in Part 3. Remove or clearly mark anything you could
not check. It may be shorter than the original. That is fine.


======================================================================
PART 6. REFLECTION (about 100 words)
Where, specifically, was your own judgment needed? Name at least one
moment where you had to decide something the sources or the assistant
did not decide for you.


======================================================================
PART 7. AI-USE RECORD
AI-USE RECORD
Label:            AI allowed — keep a record
Assistant:        (name and version, or "local model")
Date(s):
Prompts:          (paste them, or a link to the conversation)
What I kept:
What I changed:   (and why)
What I verified:  (which claims or results I checked, and against what)

======================================================================
PART 8. SELF-ASSESSMENT (do this before opening the answer key)
Score yourself honestly with the rubric in the lesson.

 Verification coverage            ___ / 30
 Error detection and correction   ___ / 30
 AI-use record                    ___ / 20
 Reflection                       ___ / 10
 Product                          ___ / 10
 TOTAL                            ___ / 100

Cap check: is the AI-use record complete?  YES / NO
(If NO, your total is capped at 60, however good the rest is.)
