Exploring Faculty Perspectives on Using AI to Support Individualized Learning Plan Goal-Writing

Competency-based medical education (CBME) has gained favor in the United States,¹ with individualized learning plans (ILPs) serving as a fundamental component to tailor developmental trajectories.^2,3 All residency specialties must now support residents in creating ILPs,⁴ a process requiring significant time and faculty development to master reflection, goal generation, and progress tracking.^2,5,6

Current best practices, gained from pediatric residencies who utilized ILPs a decade earlier, include providing residents with external feedback^6,7 and SMART (specific, measurable, attainable, realistic, timely) goal training.^8,9 Although medical student ILPs have been shown to foster self-directed learning,¹⁰ medical schools vary in whether and how they support ILPs, particularly in relation to learner motivation, specific feedback, protected time, and faculty guidance.¹¹ This inconsistency limits goal-writing proficiency of incoming interns,⁶ and it is reasonable to assume that residency programs show similar variability in how ILP processes are supported. These challenges can impact the development of meaningful, self-directed goals for residents.

Generative artificial intelligence (AI) may offer a way to assist with ILP goal-writing while increasing careful implementation of AI utilization in medical education.^12,13 No literature exists on AI use for ILP goal-writing in residency. Because faculty help guide the ILP process and faculty adoption of AI has lagged behind that of learners in higher education,¹⁴ this pilot study focused on faculty perspectives. The study objective was to assess the feasibility of using AI to transform resident learning goal ideas into SMARTIE goals (adding inclusivity and equity considerations to SMART),¹⁵ serving as a foundation for future larger-scale investigation into AI-augmented ILP goal-writing.

Eligible participants included faculty at six family medicine residency programs willing to use AI with residents during ILP goal-writing. Recruitment occurred via in-person and email communication; no incentives were offered.

Participants were asked to use AI to transform raw resident input and faculty feedback into SMARTIE goals. Given the varying ILP timelines and AI experience, participants were allowed to integrate AI into resident ILP goal-writing at their discretion. The study team provided an optional AI primer and a sample prompt (Appendix A); examples of AI integration are shown in Appendix B.

REDCap surveys were used to collect demographics and faculty post-completion perspectives. The latter included five open-ended, exploratory questions (no validity evidence; see Appendix C) regarding AI utility for ILP goal-writing. Given the small number of responses, an exploratory qualitative approach was used to synthesize comments into key categories. Two authors independently reviewed the data to ensure reliable interpretations.

This study received exempt status by each investigator’s institutional review board. Informed consent was obtained via REDCap prior to collecting any data.

Eleven faculty participated; 63% had been an advisor for five years or less. Table 1 summarizes the demographics collected. Of the participants, 64% completed the open-ended questions.

Review of responses to the open-ended survey questions identified four categories (see Table 2 for example responses).

1. Feasibility of AI Support: Faculty found integration of AI to be feasible and helpful, particularly when residents had difficulty translating abstract ideas into actionable steps. Faculty described AI-assisted goals as more structured, specific, and aligned with competency frameworks. AI support was especially helpful when trying to phrase goals in the SMARTIE format.

2. Faculty Satisfaction with AI Support: Faculty reported that using AI reduced time spent helping residents in ILP goal-writing. Several found AI-generated goals clearer and more actionable than those in initial drafts. Additionally, several reported that the process made them less hesitant to use AI in the future. Comments suggest that once exposed to AI, faculty were more predisposed to trust it as a tool, provided there was appropriate human oversight.

3. Approaches Used for Integrating AI Support: AI approaches used for ILP goals varied; see Table 3 for examples. AI was used to refine resident-generated drafts or to translate jointly discussed targets into specific tasks. AI also streamlined goal refinement but required manual adjustments for personalization.

4. Challenges of AI Support: Challenges identified by faculty included inconsistent AI literacy, ethical concerns around privacy, and uncertainty about whether AI-assisted goals were truly resident-driven. Additionally, refining prompts to achieve optimal goal output was noted to be a hurdle. Balancing protecting personal information with providing enough personalization to achieve effective results was also challenging.

Faculty feedback in this pilot study suggested that using AI to assist with resident ILP goal-writing was feasible. However, faculty clarified that developing appropriate prompts was key to deriving more specific and appropriate learning goals from using AI.^16,17 Detailed AI prompts were deemed helpful in creating learning goals that better adhered to SMARTIE principles.^6,7 While faculty felt that AI augmented the ILP goal-writing process in an efficient and convenient way, they did not see AI as a replacement for experience or insight that can be shared during the resident-faculty ILP interactions.¹⁸ The use of AI in this manner may thus be helpful when residents or faculty have not attained skills in goal-writing, or when programs do not have optimal time or levels of support for CBME.^8,19

The small number of participants in this pilot limits generalizability. Program-specific variability in ILP timelines and processes reduces study reproducibility. The study also lacked direct resident input on AI utility for drafting ILP goals, and the cross-sectional nature of the data collection was process-oriented, without seeking evaluation of outcomes or resident goal achievement.

Faculty perceived that AI was a feasible tool to assist with resident ILP goal-writing, yielding relative improvement in creating SMARTIE goals. With a carefully developed prompt, AI could quickly enhance learning goals by improving their specificity and structure, although faculty still viewed their own input to be helpful for providing guidance and ensuring accuracy and practicality. These preliminary findings suggest AI may reduce the cognitive load and time constraints associated with generating high-quality, more specific learning goals, particularly through structured templates supported by carefully developed prompts grounded in established frameworks. Future research is planned to examine standardized prompts and the impact of AI-augmented ILP goals on resident progress.

Preliminary data from this pilot study was presented via a Works in Progress poster at the May 2025 Society of Teachers of Family Medicine Annual Conference in Salt Lake City, UT. The pilot was also shared in the Innovation Showcase at the 2026 Residency Leadership Summit in Dallas, TX.

The authors have no conflicts of interest to disclose.