Session 4 · Student worksheets
← All sessions · Download Word workbook
Type answers below, download a text copy, or print this page. Drafts are saved only in this browser when browser storage is available. They are not submitted to the instructor. Avoid shared devices for private answers.
พิมพ์คำตอบในช่องด้านล่าง แล้วดาวน์โหลดเก็บไว้ได้ คำตอบยังไม่ได้ส่งให้ผู้สอน ส่วน exit ticket ให้ส่งผ่านลิงก์ท้ายใบงาน
Worksheet 4A AI output audit
28 October 2026 • 10 minutes
Evidence: at 350 °C, mean conversion is 64% for baseline and 75% for modified catalyst, with three independent batches per group. Evidence covers one feed and an 8 h window. No activation-energy study or industrial lifetime data exist. All values are synthetic.
Mock AI output for critique
“The modified catalyst improved conversion by 75% and performed better under all industrial conditions. This proves a lower activation energy and a long commercial lifetime. These findings agree with REF-X.”
REF-X is an intentionally invented placeholder. This paragraph is a teaching example written to contain errors, not a measured evaluation of any AI model.
| Claim or change | Evidence check | Accept revise or reject |
|---|---|---|
Corrected paragraph
Verification record
Which checks did you perform yourself?
Worksheet 4B Ethics and disclosure
Explain the action and reason. Use the journal and institutional policies that apply.
1. A language-editing tool invents a reference in your manuscript.
2. A collaborator plans to upload an unpublished industrial dataset to an unapproved AI service.
3. An image generator produces a realistic SEM image for use as experimental evidence.
AI use log
Draft disclosure
Describe actual use only. Identify the tool, purpose and human review. Verify the required wording and location for the target journal.
Policy example checked 6 October 2026: Elsevier generative AI policies for journals. Read the current page before submission: