Can We Use AI to Assist with RE Curriculum Development? 10 Practical Considerations

The Bottom-line at the Top: Artificial Intelligence (AI) is being integrated into every profession. Can relationship educators use AI to help with their curriculum development needs and provide a high-quality product? In this article, a group RE practitioners share their pilot-testing journey using AI to help develop new e-courses. They learned that AI assistance can significantly shorten the development time required for new e-courses and even help improve curriculum in some ways. But they argue there is a need for some best practices, and they alert developers to key considerations and concerns. AI still needs knowledgeable and careful human oversight when assisting with RE curriculum development.   

Citation: Clyde, T. L., Hawkins, A. J., Higginbotham, B. J., Gordon, R. S., & Schramm, D. G. (20260. Practical considerations for AI assistance with relationship education curriculum development. Family Relations. https://onlinelibrary.wiley.com/doi/10.1111/fare.70197 

Study Set-up

• Virtually every knowledge-based profession today is experimenting with how Artificial Intelligence (AI) can increase work efficiency and improve work quality. The relationship education (RE) field is no different. 

• Relationship educators are beginning to explore how AI can assist with needed curriculum development. But currently there are no guidelines on how to do this effectively and ethically, and there are no lists of practical considerations when using AI. 

• Drawing from their work designing relationship education (RE) e-courses for adult learners, a group of RE curriculum developers have learned firsthand what AI tools can do but also confronted new pedagogical questions and ethical considerations. Some of these are the result of current weaknesses with AI.

Study Background

• The Utah Marriage Commission is an in-statute Commission charged by the Utah legislature to strengthen the institution of marriage and help couples form and sustain healthy relationships and stronger marriages. The Commission is hosted by Utah State University Extension with its mission to “take the research to the people.”

• On its website, StrongerMarriage.org, the Commission provides a wide variety educational resources, including free (for Utahns) e-courses. The Commission has developed several e-courses the “old-fashioned” way – with human curriculum developers. As relationship educators know well, this is a labor-intensive task but yields a high-quality product. 

• In 2024, the Commission’s curriculum development team decided to experiment with AI assistance in this curriculum development process. Could AI help to create high-quality e-courses and reduce the time and resources needed? What are the possibilities and pitfalls? 

• What did they learn? And what do they recommend for the RE field? The article summarizes this learning journey in 10 recommendations for the field. 

10 Practical Considerations

Preparation for Curriculum Development

1. Invest in learning how to work with AI. First, get literate with AI’s capabilities, constraints, and weaknesses. Get training in AI tools and their effective use. Don’t just wing it. (These curriculum developers decided to use ChatGPT and ScholarGPT.)

2. Ground yourself in the relevant research. Curriculum developers should invest time in the traditional academic and human process of reviewing and synthesizing relevant research and scholarship on the topic. AI can assist human intelligence, but it needs to be a collaboration. And a solid foundation in the research allows the human curriculum developer to take the lead in what content is most important to cover. Moreover, it allows them to spot inaccurate or even fabricated information often provided by AI tools. Humans should still lead out in development. AI is best at summarizing the specified content and creating learning activities. 

3. Use AI to scope out the questions people are asking about the topic. AI can quickly scope out the web to find out the most common questions people are asking about the topic. This can drive curriculum content to be more responsive to users’ needs and interests. Curriculum developers can then match good research to the questions. 

4. Give AI clear instructions for your project and invite clarifying questions. Human curriculum developers should give AI detailed instructions on what they want for the course. For instance, tell AI who your target audience is. Are there any sensitivities for this audience? What format do you want? How many lessons? How long are the lessons? What learning techniques do you want to use (e.g., couple discussions, self-exploration guides, min-quizzes, etc.)? What tone do you want (e.g., formal-serious, casual-conversational, first-person voice, second-person voice, etc.)? Then, invite AI to ask clarifying questions to increase accurate responses to your instructions.
   
   Assistance with Curriculum Development

5. Instruct AI to create content based on good scholarship. Diligently train AI to draw from scientific, scholarly sources, not just scrape the web. Developers should consider feeding AI specific studies that are key to the topic. Note: AI may not be able to get past paywalls to get to some of the best research. And AI is not yet adept at discerning between stronger and weaker research. Developers need to vigilantly check for AI-generated “hallucinations. (AI is really good at inventing fictional research!)  

6. Use structured prompts to guide AI assistance. Guide AI assistance with effective prompts. And break the educational goal into distinct parts. For instance, give prompts such as:
   - “Review this study and use it as a primary source for this section.”
   - “Write a realistic, inclusive, and relatable narrative, about 2-3 paragraphs in length, of a fictional couple experiencing the problem discussed in this section.”
   - “Review the content of this lesson and generate three review quiz questions (true or false) on key points at the end.”  

7. Refine and revise. Human curriculum developers should use their knowledge and expertise to review all sections and the overall course (or invite those with greater expertise to review). Then refine the content. Consider using prompts like:
   - “Can you make this paragraph less wordy but keep the warmth?”
   - “Make this narrative feel ‘messier’ – like a real couple conversation, not a scripted dialogue.”
   - “Incorporate the key findings of this specific study in this section.” 

8. Stay in the same chat; don’t start new ones. This preserves continuity and context. It retains instructions on writing style, content structure, and human feedback from previous “chats” with AI.

9. Do a final integrative and human review. AI is not integrative in the human sense. Do a final integrative review to refine and polish. Review the overall quality of the research supporting the content. And check for plagiarism and appropriate citations. Also, AI can be quite repetitious across sections in terms of words and phrases used; eliminate unnecessary repetition. 

10. Disclose your use of AI assistance. The level of detail developers should provide to the user is open for debate and may depend on context. But some basic disclosure is both ethical and provides some transparency that increases trust. Here is an example:
    - “During the preparation of this work the authors used GPT-4o and DeepResearch to identify and summarize scholarly research as well as refine content language, including but not limited to tone, clarity, appeal, and search-engine optimization. After using GPT-4o and DeepResearch, the authors reviewed and edited the content as needed. The authors take full responsibility for the published content.”

Implications for the RE Field

• Can RE developers use AI to reduce the time and effort needed to develop content? Yes, it can save time in the long run and produce good content. And it does some things well, such as generate fictional and relatable narratives and create engaging activities. But you have to know how to use AI effectively and invest time in becoming an expert in the content. And you need to direct AI’s work carefully and always check, review, and revise. You shouldn’t just turn over the task to AI. Humans lead; AI assists. 

• Bonus Implication: As AI gets better, will it replace the need for human content developers? Doubtful. RE is a very human process and the human touch will always be needed. Plus, a knowledgeable human will need to set the parameters for learning.  

Last Word: AI can assist in developing RE curricula. It’s good at doing some things that will save you time, but it needs human intelligence to oversee and refine its work. AI can be an effective assistant, but it shouldn’t be leading the curriculum development process.