AI Literacy for Psychotherapy Educators: 8 Questions Before Adoption

Conversations about AI in therapy education often begin with a feature list: “Can it role-play? Can it give feedback?”

For educators, a more important starting point is: “Which learning problem are we trying to solve, and how will we recognize meaningful progress?”

AI literacy is not simply the ability to write prompts. It includes understanding capabilities, limitations and responsibilities when using AI for teaching.

Eight questions worth asking

1. Which exact skill are we targeting? “Make therapists generally better” is too broad. Start with something observable, such as emotional reflection.

2. What does the tool do—and not do? A system may generate plausible dialogue without being able to assess therapeutic relationships or readiness for clinical work.

3. Who designs the scenarios? Ask how virtual characters, difficulty levels and educational appropriateness are reviewed.

4. How is feedback produced? Are criteria visible? Can a supervisor inspect the relevant excerpt and disagree with an automated suggestion?

5. What happens to data? Review collection, retention, access, vendors and deletion processes. Entirely fictional scenarios are often a safer starting point than real client material.

6. How are errors and bias handled? Persuasive AI feedback can still be inaccurate. Systems may also favour particular linguistic styles or cultural norms.

7. Where does human supervision enter? Supervision should not become a rubber stamp for an automated score. Educators need access to learning objectives, dialogue evidence and the rationale for feedback.

8. How will educational value be assessed? Do not stop at usage hours or learner satisfaction. Plan comparable baseline and post-training tasks and document limitations.

Why does this matter now?

The American Psychological Association's 2025 ethical guidance discusses transparency, bias, data security, misinformation and human oversight in AI use within professional psychology.

In the European Union, Article 4 of the AI Act requires providers and deployers to take measures supporting AI literacy among relevant staff and others operating systems on their behalf. What this means for a particular organization depends on its role and context. It does not imply one universal AI qualification or identical requirements for every educational tool.

A useful faculty exercise

Before choosing a platform, select one fictional training scenario. Ask two supervisors to agree on what an effective response would look like. Only then compare their criteria with an AI system's suggestions. Disagreements can reveal important decisions for curriculum design.

From theory to practice: AI earns its place in a curriculum by serving a learning objective—not simply by being available.