Several population groups continue to be underrepresented among trainees pursuing biomedical research disciplines [1-5]. According to 2023 census data, Hispanic, Black, and American Indian and Alaska Native individuals comprised 34.5% of the US population (19.5%, 13.7%, and 1.3%, respectively) [6]. Yet, only 16.0% of students awarded science and engineering doctorate degrees in 2020 were from these groups [7]. These disparities carry over into the workforce, with Hispanic individuals comprising only 15.0% of the STEM (science, technology, engineering, and math) workforce in 2021, 9.0% of Black individuals, and <1.0% of American Indians and Alaska Native individuals [7]. These figures illustrate a disparity between the racial and ethnic diversity of the US population and of individuals working in biomedical and STEM fields.

The underrepresentation of trainees from diverse backgrounds is particularly pronounced in the fields of biomedical artificial intelligence (AI) and machine learning [8,9]. Biomedical research has seen significant growth in AI applications with the development of tools for diagnosing, monitoring, and predicting prognoses [10,11]. With AI’s increasing role in society, AI researchers’ diversity must mirror the diverse populations they serve. A skilled research workforce in domain-specific AI applications is needed to maintain sustained progress in AI research. According to a 2019 report from the AI Now Institute, women comprised only 18% of authors at AI conferences, while 80% of AI professors were male [8]. A lack of diverse perspectives can have significant consequences on resulting research products, as lack of fairness (the ability of AI to be free from bias) and explainability (the ability to understand why an AI model makes certain predictions, which is crucial for identifying biases) have plagued the AI field for years [10,11].

Common barriers to students’ participation and persistence in research include experiencing feelings of exclusion and a lack of belonging, difficulties developing a strong research identity, perceptions of racial and social stigma directed toward them, and perceived gaps in research skills [1-3,12,13]. One study, for example, observed that reasons for dropout from computer science degrees included unclear expectations about the chosen field, feeling as if they did not belong, and social isolation [12]. These challenges are particularly pronounced among trainees from groups traditionally underrepresented in research, as articulated by the National Institutes of Health (NIH) [14]. The NIH reports that several racial and ethnical groups are underrepresented in research: Black or African American individuals, Hispanic or Latinx individuals, American Indians or Alaska Natives, and Native Hawaiians and other Pacific Islanders, with women carrying a higher burden of underrepresentation [14]. To address these known barriers, research exposure programs for undergraduate students can include a curriculum of academic skill development components that promote student self-efficacy, identity, and a sense of belonging in research. Specifically, evidence shows that students encouraged to develop their “science identity” are more likely to integrate into the wider culture of scientific research [3]. Students can develop their self-confidence, self-efficacy, and identity as scientists through science support experiences and connections with peers and mentors, thus benefiting from exposure to diverse perspectives and experiences [5,15,16]. Additionally, experiences that foster a sense of belonging to the research community have been shown to lead to greater persistence in STEM fields, especially for students from groups underrepresented in research [16-19]. Furthermore, STEM undergraduates who engage in small-group and active learning demonstrate increased self-efficacy, achieve better grades, and are more likely to remain in STEM fields [16,20,21].

There is a critical need for nationwide collaborative training programs to fund, develop, and mentor students from underrepresented groups to ensure AI representation in biomedical research [22]. Summer research programs delivered via both in-person and web-based formats have been shown to increase the participation and retention of undergraduate students in research [23]. Evidence shows that involvement in both face-to-face and remote summer research experiences equips undergraduate students with knowledge and research skills and promotes enrollment in postgraduate training and entry into scientific- or health care–related careers [24,25]. In addition to being exposed to research and learning research skills, students can also gain confidence as scientists and develop a researcher identity and intention to pursue a career in STEM [19,26,27]. However, previous programs have primarily focused on equipping students with technical knowledge and skills, with varying degrees of emphasis on professional development, without integrating a universal academic enrichment curriculum that is delivered synchronously via the web to students across multiple sites and that focuses on active learning and improving students’ academic and research skills [23-27]. The effective development of students’ research skills and familiarity with scientific methods calls for supplementing laboratory research experience with academic enrichment sessions in order to equip students with the necessary academic and professional skills to thrive as researchers building their identities in the research space. These sessions should allow students to actively engage in discussions about their identities as researchers while offering a supportive environment that fosters a sense of belonging and facilitates peer-to-peer information exchange. This study aimed to report the findings of delivering academic enrichment sessions to undergraduate students from groups underrepresented in AI research and to test the hypothesis that a guided academic enrichment curriculum, as part of a summer research internship that also included a laboratory-based experience, promotes self-perceived research skills, researcher identity, and professional skills.

This paper reports on the pilot implementation of an academic enrichment curriculum delivered as part of the summer, AI-focused, undergraduate research internship administered through the HuBMAP program. Following participation in the internship, students reported increases in self-assessed research skills, including their abilities to set research career goals, to think of themselves as researchers, and to work comfortably in the research environment, and increases in their feelings of belonging in research. Our curricular intervention included academic enrichment content supplemented by peer learning and hands-on activities.

Earlier research has demonstrated that perceived gaps in research skills and feelings of isolation have detrimental effects on students’ persistence in computer science and AI-focused training [12]. While most summer research internships focus on individual training and technical aspects of research with varying degrees of emphasis on professional development [23-27], the format and content of the supplemental academic enrichment curriculum we piloted showed promise in creating a nurturing, active, and collaborative learning environment for student interns across multiple institutions, with demonstrated self-reported improvements in abilities that went beyond technical research skills.

Based on our study’s findings, we have formulated propositions and hypotheses about the mechanisms through which academic enrichment curricula promote research skills self-efficacy and researcher identity and result in sustained engagement in research training. First, delivering an academic enrichment curriculum enhances and amplifies the learning experiences of students participating in summer research programs and promotes the long-term retention of these students in biomedical research careers. Second, academic enrichment sessions help to build a research community and promote a sense of belonging among students. Third, providing students with opportunities to learn from one another encourages peer learning and enhances students’ critical thinking skills, resourcefulness, and resilience within academia. Additional research questions that we would like to answer through future studies and longitudinal assessments include how students can learn from one another and whether peer learning is superior to learning from mentors. Future studies should also explore the effect of individual versus cohort-based research experiences. The academic enrichment curriculum we delivered to participants of the HuBMAP summer internship program would be considered a mix of cohort (delivered through the academic enrichment sessions) and hybrid (since students were based at separate laboratories and institutions with individual mentors) research experiences. In addition, further research is needed to determine the optimal structure of the academic enrichment curriculum, including the frequency of sessions (once weekly vs every other week, for example).

Our experience suggests that the integration of an academic enrichment curriculum into a laboratory-based AI-focused summer research internship can help students increase their confidence, communication skills, self-efficacy, and sense of belonging in the research community. In line with our findings, a growing body of evidence shows that a structured curriculum supports undergraduate students from traditionally underrepresented groups in developing their identity as scientists, and in their overall self-efficacy to use scientific thinking, through science-support experiences and connections formed with peers [5,15,16]. Thus, we recommend that future programs should incorporate small group peer learning, active learning strategies, and laboratory-based research experience as evidence-supported methods to facilitate increased self-efficacy and a sense of belonging in the research community, which may also have the potential to promote the retention of undergraduate students in STEM and biomedical disciplines [16,20,21].

Despite these encouraging results, it is important to recognize that the generalizability of the results of this study is limited by its relatively small sample size of 22 students, of which 14 students (63.6%) completed the pre- and postprogram surveys. As a result, this study’s findings may not fully represent the broader experiences of other students across other programs. It is also important to note that it was not possible to separate the effect of our academic enrichment curriculum from other activities that students may have been engaged in during their HuBMAP Summer Research Internship. Multiple individual laboratories across different institutions hosted students participating in the 2023 HuBMAP summer internship program. Therefore, while we delivered this academic enrichment curriculum to all students via the web, some students may have been provided additional learning opportunities (remote or in-person) at the laboratories or institutions where they were based. The nature and quantity of such other learning opportunities may have varied across the different sites and should be explored in future studies. It is possible that a portion of the gains in students’ self-reported research abilities were attributable to these supplementary activities and not because of the academic enrichment curriculum. Future research should therefore investigate the specific impact of academic enrichment curricula on the long-term retention of students in biomedical research careers, particularly retention of students underrepresented in biomedical fields. Furthermore, each student received one-to-one and group mentoring from lead investigators and senior trainees of the host labs. The effects of mentoring and the role of peer-to-peer interaction as contributing factors in student academic skill development can also be quantified and measured in future evaluations of summer research experiences. Additionally, this study used a pre-then-post evaluation design, which combined the questions in which students were asked to rate their research abilities “before” and “as a result of” the HuBMAP Summer Research Internship into a single survey administered post program. This methodological design allowed the students to assess knowledge gained from a perspective informed by a 10-week experience. However, the effects of recall and social desirability biases can be evaluated with sufficiently powered studies that use different evaluation formats.

Following their participation in the HuBMAP 2023 summer internship program, which included an academic enrichment curriculum delivered remotely, students’ self-reported research abilities, including their confidence and communication skills, their self-efficacy in research, and their abilities to set research career goals, increased. Students also reported an increased ability to think of themselves as researchers and to work comfortably in the research environment, as well as increased feelings of belonging in research after participating in the program. We recommend that summer internship programs incorporate an academic enrichment curriculum with small-group peer learning in addition to a laboratory-based experience to facilitate increased student engagement, self-efficacy, and a sense of belonging in the research community. Strengthening the pipeline of skilled and diverse professionals entering these fields will help in addressing both current and future workforce needs, reducing bias in AI fields, and cultivating the next generation of scientists and leaders in public health, with the goal of improving health outcomes for all populations.