By Cody Saunders, LMSW · October 11, 2026
Program abstract
A care team needs more than a rule that says “use AI responsibly.” Staff need to know which tasks are allowed, what information may go into a tool, who checks the work, and where to get help. Clear rules can make daily choices easier and give useful tools a fair chance to help.
This free talk helps behavioral health leaders turn broad goals into steps staff can follow. We start with one task, such as making a staff handout from an approved public policy. We then compare it with a care task, such as drafting a visit note. Both involve writing, but the data, review, and effect on people differ.
Participants learn how to describe an approved use, assign a reviewer, and explain when to stop. A made-up team example shows why the account, feature, and information matter as much as the product name. The session ends with a short rule participants can take back for review by their own team. The goal is less guesswork and a clearer path to useful work, with more room for care and human connection.
Audience and format
This proposed 40-minute talk is for practice owners, team leads, supervisors, and staff who help set AI rules. Clinicians, privacy staff, and IT staff can join together. No coding skills are needed. It includes a short exercise and questions. It is offered as free speaking. It does not promise continuing education credit or certify a team’s policy.
Before the talk, the host can share the audience size, work setting, and learning goals. Use public or made-up examples. Do not send client records for the session. The talk teaches a way to write clear rules; it does not approve a tool or replace the host’s legal, privacy, or clinical review.
Learning objectives
By the end, participants should be able to:
- Describe an AI use by task, account, feature, information, and purpose.
- Explain the difference between a law, a professional standard, and a team policy.
- Write a short rule that names what is allowed, who reviews it, and when to pause.
- Name one way to test whether a rule is clear and workable for staff.
Proposed session flow
0–5 minutes: Start with the work. Ask what daily task feels hard or slow. Use an example that involves no private data. Explain that faster drafting is useful only if the final work remains accurate and the full effort is manageable.
5–12 minutes: Make the use specific. Compare a public staff handout with a visit-note draft. Name what goes in, where it goes, and who uses the result. A familiar app may send information to another service. Approval for one feature does not automatically cover a new recorder or summary feature.
12–20 minutes: Turn a broad rule into a clear one. Show the sample below. Assign an owner for questions, an approved account, a source, and a reviewer. Separate permission to draft from permission to finalize. Explain how staff can report a problem without having to solve it alone.
20–28 minutes: Practice with a made-up team. Participants read the case and write one sentence about what must change before use. Invite two different roles to compare their answers.
28–35 minutes: Check the rule in daily work. Ask someone who did not write it to explain the next step. Look for missing details. Plan a small trial with approved information. Count review time, errors, and staff effort as well as draft time. Set a date to revisit the rule.
35–40 minutes: Questions and one next step. Each participant chooses one use to describe more clearly. Encourage the host to bring that description to the people who own the relevant decisions.
A sample rule to discuss
“For this trial, trained staff may use the approved work account to turn the current public staff policy into a draft handout. Do not enter client information. The named policy owner compares every step with the source and approves the final handout. Stop and contact the trial lead if the tool adds a rule, requests private data, or changes its data practices.”
This is a teaching example, not a ready-made policy for every team. Before adopting it, specify the actual account, policy version, owner, and help contact. A rule should tell staff what to do when the approved tool is unavailable. Otherwise, time pressure may lead to an unapproved shortcut.
Made-up case and answer notes
A clinic approves a tool for public staff handouts. A new feature offers to record visits and draft notes. A staff member says it must be allowed because the product is already approved. The tool can save the recording, but the clinic has not reviewed where it goes or how long it stays.
Ask: What should pause? Who needs to review the change? What can staff still do under the existing rule?
Debrief: Pause the new visit-recording use. Keep any existing approved use within its limits. The new use adds private information and care-record work. Clinical, privacy, security, and other responsible staff need to review the exact feature and data path. Name the final reviewer and the patient explanation and consent process that applies. Do not treat a product-wide approval as permission for every feature. Do not ask participants to diagnose a legal violation from this short case.
A shared framework for decisions
Use the existing Behavioral Health Responsible AI Framework rather than adding a second framework. Its five domains are Clinical Appropriateness, Privacy & Security, Reliability & Safety, Human Oversight, and Governance & Accountability. They help teams record the evidence, people, and limits behind a decision. The framework is evolving; it is not a validated score or certification.
Sources and teaching scope
Sources checked October 11, 2026. The flow, rule, and case are original teaching examples. They do not establish that a tool lowers costs, improves billing, or improves care. Those possible benefits need to be measured in the actual workflow.
NIST’s AI Risk Management Framework provides voluntary guidance for managing AI risks. It is not a law or a product approval. NIST notes that version 1.0 is being revised. This talk draws on the need to assign roles and review use over time; the sample rule is Cody’s teaching proposal.
HHS guidance on HIPAA and cloud computing explains duties for covered organizations and business associates handling protected health information. A business associate agreement and other safeguards may be required for the proposed use. A team policy cannot replace applicable legal duties. The host’s responsible review team must determine the rules for its own setting.
Resources to share
Use the purchasing checklist and data-handling worksheet to support a specific review. Read the governance article for a fuller team process and AI Terms for Care Teams for plain-language definitions.
To discuss a free talk for your organization, visit Work With Me.