Pre-Conference- AI My Prevention Work
This course is also listed on the MCTC website as “How to Leverage AI in Prevention” 3.0 IBC CEUs have been requested.
Session Description: Most prevention specialists have heard plenty about AI and have been handed very little they can use on a Thursday afternoon. This training is the opposite. It is built for people who have never used an AI tool at work or tried one once and stopped.
Participants leave having built two things: an action list with owners and dates, drawn from a coalition meeting transcript; and their choice of parent-facing flyer copy drafted from an approved source packet, or a reviewer’s flag list on a grant narrative. They also leave with a one-week plan for one workflow in their own coalition, with a named reviewer.
The training is tool-agnostic. A free ChatGPT, Claude, or Gemini account works, and so does a county-licensed tool. Participants bring a laptop if they can; two people can share one. All practice material comes from a fictional Iowa coalition, so nothing real or sensitive goes near a chat tool.
AI should support prevention work, not replace expert judgment.
Training goal
Participants leave able to run one low-risk, human-reviewed AI workflow in their own prevention work within one week and knowing what must never be entered into a chat tool.
SMART learning objectives
By the end of the 180-minute session, participants will be able to do each of the following. Every objective is demonstrated in the room, and the evidence is the participant’s own completed workbook.
| # | Objective | Evidence of achievement |
| 1 | Write a four-part prompt — goal, context, source, expectations — for a routine prevention task, scoring at least 3 of 4 on the prompt checklist. | Workbook Checkpoint 1, self-scored 0–4 against the printed rubric. |
| 2 | Produce a decisions-and-action-items list from a coalition meeting transcript and verify at least one owner and one due date against the source document. | Product 1 completed and one verified fact recorded in the workbook; checked against the answer key during debrief. |
| 3 | Produce either parent-facing flyer copy drafted from an approved source packet, or a reviewer’s flag list on a grant-narrative draft — including a written AI-disclosure line. | Product 2 completed in the participant’s chosen track; disclosure line written in the workbook. |
| 4 | Diagnose one unsatisfactory AI output by identifying which part of the prompt caused it, and revise one variable at a time in a follow-up prompt. | Diagnose-and-revise step recorded in the workbook during Product 2. |
| 5 | Assign a risk tier — lower, moderate, or higher — to their own mapped workflow, and name at least three categories of material that must never be entered into a chat tool, including youth-identifying information and survey data below the aggregate level. | Workflow map with risk tier marked; keep-out list completed; exit ticket. |
| 6 | Complete a one-week pilot plan naming the workflow, the owner, the source materials, the reviewer, and the success measure. | Pilot plan page in the workbook with all five fields filled in. |