NSPA 2026 · October 21 · 3:30–5:00 PM CT
Trust, Transparency,
and AI
Building Responsible Scholarship Review Practices
Build a review prompt, map a protected workflow, and draft a fair approach to applicant AI use.

The 90-minute workshop agenda
| Time (CT) | Focus | What you make |
|---|---|---|
| 3:30–3:38 8 min | Frame the decisions | Identify one review task and rate a starting capacity. |
| 3:38–3:50 12 min | Build a structured prompt | Draft a constrained prompt with an evidence requirement. |
| 3:50–4:05 15 min | Demonstrate administrative review | Test synthetic application triage; catch omissions and unsupported claims. |
| 4:05–4:20 15 min | Calibrate a review rubric | Compare independent human ratings before seeing an AI draft. |
| 4:20–4:30 10 min | Protect the workflow | Place a data gate and a named human decision owner in a workflow. |
| 4:30–4:45 15 min | Evaluate applicant AI use | Interpret a detector signal; draft a fair policy response. |
| 4:45–4:55 10 min | Choose a capacity target | Select one capacity, evidence of progress, and a 30-day action. |
| 4:55–5:00 5 min | Commit and close | Share an exit ticket and address remaining questions. |
Use during the session. Keep afterward.
Your workshop materials
Download the offline webdeck · Slides and notes work offline; linked online labs need an internet connection.
Quick references as online images
Browse all 26 individual slide images
01. Trust, Transparency, and AI
02. Three things to take home
03. People make the decisions
04. Where are you today?
05. Give the AI a clear job
06. Remove the guesswork
07. Activity: improve your request
08. Meet our practice applicant
09. Demo: organize, then check
10. What can we support?
11. Activity: check one row
12. Use the same scoring guide
13. Compare your reasons first
14. Demo: compare the AI draft
15. Before using real information
16. Activity: name the checkpoints
17. A detector flag is not proof
18. Small error rates affect people
19. Make the rules clear before people apply
20. Activity: ask without assuming
21. Choose a skill to work on
22. Activity: mark, choose, and plan
23. Try one small change this month
24. What will you take back?
25. Keep the materials close
26. Sources and scopeSources, interpretation, and reuse
Reviewed September 24, 2026. Recheck product guidance before the October workshop. The packet, rubric, policy starter, and capacity matrix are workshop teaching tools, not validated selection instruments or official NSPA policy.
- NIST Gen AI Profile (2024)
Gen AI can produce confident falsehoods; evaluate performance in the intended setting, verify sources, and document oversight.
- Turnitin detection FAQ (reviewed September 24, 2026)
The indicator estimates the share of qualifying prose flagged by the model. It is not a probability of misconduct or a determination of misconduct; Turnitin advises against using it alone for adverse action.
- Liang et al., Patterns (2023)
This study found false-positive disparities for the non-native English writing and detectors it tested. It does not establish error rates for all current tools.
- Al Ali, Helcl & Libovicky, EACL SRW (2026)
A later study centered on Czech writing found no systematic non-native-speaker bias across its tested detector families. Context, language, and detector matter.
All examples are fictional. Use an organization-approved workflow before processing any real applicant data. Policy language needs local review. Generated artwork supports the slides; all instructional wording is separately typeset.

Prompt & rubric
Protected workflow
Fair applicant AI use
Capacity matrix · page 1
Capacity matrix · page 2
Capacity matrix · page 3