According to CASEL, a leader in the field of social emotional learning, “SEL is the process through which all young people and adults acquire and apply the knowledge, skills, and attitudes to develop healthy identities, manage emotions and achieve personal and collective goals, feel and show empathy for others, establish and maintain supportive relationships, and make responsible and caring decisions.” This paper intends to explore how the emerging field of Artificial Intelligence has contributed to SEL in the recent past and its potential impact in the future on this research-proven practice in education. As McStay explains “schools around the world have relatively limited resources for teaching and professional development related to social and emotional learning. As such, one can see the appeal of technologies that promise intimacy-at-scale to solve a longstanding problem in education: how to provide personalized attention to packed classrooms” (McStay, 2019, p. 9).

Collaborative for Academic, Social, and Emotional Learning (2024)

This topic is important since SEL has been proven to have a positive effect on “students' social, emotional, behavioral, and academic outcomes at all grade levels and across gender, ethnicity and race, income, and other demographic variables.” Research has consistently shown that people with high EI (emotional intelligence) have leadership potential, good decision-making, and interpersonal skills, which can result in good mental health. Social, emotional, noncognitive, and soft skills are often defined as skills that are susceptible to interventions and policy measures, especially during the early years of an individual’s life (Chernyshenko et al., 2018). By integrating human EI principles into the AI design of machines, developers are able to produce technologies that are more intuitive, sympathetic, and user-friendly (Luckin, 2024).

As AI systems continue to develop, exploring the relationship between AI and SEL can help us gain a deeper understanding of the capabilities and limitations of such a promising partnership. Although the concept of AI has been around for many decades, it is a new discipline in the field of education with most research being conducted in the last 20 years. AI-driven systems incorporate “empathetic algorithms” that are capable of “recognizing, interpreting, and responding to human emotions in a manner that fosters emotional growth and understanding” (Velagaleti et al., 2024, p. 2051). These algorithms combine cognitive computing with EI to promote deeper interactions between humans and machines. As Singh et al. explain, “in order to achieve a future where emotional intelligence (EI) and artificial intelligence (AI) coexist harmoniously, a shift in perspective on technology is essential. Technology should not be viewed as a replacement for human interaction, but rather as a powerful tool to augment and enhance our capabilities (Singh et al., 2024, p. 47).

Effectiveness of EI algorithms in facilitating positive human-AI interactions (Velagaletie et al., 2024)

The goal is to get AI-driven tools, e.g. robots, to understand and respond to the human users’ emotions in order to make the interactions more natural to meet the humans’ needs. In the field of education, AI can facilitate personalized learning as well as student collaboration (Luckin et al., 2016). Educational technology is also changing the teachers’ roles from providers of information to facilitators and supporters of learning activities, such as collaborations, discussions, and reflections (Popenici & Kerr, 2017). My experience with using AI-powered tools in the field of SEL is very limited. As a math tutor in an underserved area, I was exposed to some basic AI learning tools that students were using on the school-provided laptops. These programs incorporated very few SEL features e.g. motivations/suggestions to reduce anxiety if the student's answer was incorrect.

In the context of Artificial Intelligence, sociocultural theory refers to the idea that AI systems are created based on the social and cultural contexts in which they are created and utilized. This perspective draws on Lev Vygotsky’s idea of using “tools” to connect learning with cognitive factors. His principle of “scaffolding” includes the integration of AI-enables tools as part of learning environments. Vygotsky believed learning only takes place in interactions with an educator, parent, peer, and or guide but not individually. In the case of AI, students can collaborate with AI technology as a tool to solve a problem. Vygotsky considered the use of tools as something specifically human. As Taber explains “humans can use tools to make and improve other tools, and Vygotsky thought that this second-level use of tools was important to our development” (Taber, 2020, p. 280). These learners “need to be provided with support in their ZPD, and if the usual means of mediation were not accessible, alternatives needed to be found or developed (Taber, 2020, p. 288). Regarding the related theory of social-constructivism, Salas states that the “most recent perspective of social-constructivism indicates not only that the intellectual aspect matters, but that the physical and social-emotional aspects also play a significant role in the integral development of human beings (Salas, 2020, p. 1810).

There are criticisms of the above theories posited by researchers like Paul Kirschner, John Sweller, and Richard Clark. They explained that “because students learn so little from a constructivist approach, most teachers who attempt to implement classroom-based constructivist instruction end up providing students with considerable guidance” (Kirschner et al., 2006, p. 79). Critics also believe that exploring a complex learning environment, much like new AI tools, may ruin a learner’s experience resulting in bad educational outcomes.

To navigate this new lanscape in education, Velagaleti et al. explain that “the integration of emotional intelligence into AI systems necessitates a multidisciplinary approach, combining insights from psychology, cognitive science, and computer science. One proposed framework involves the development of a layered architecture for AI systems, where the base layer handles the fundamental emotional recognition tasks (e.g., detecting emotional cues from voice or facial expressions), the middle layer interprets these cues within context (considering factors such as culture and individual differences), and the top layer manages the appropriate emotional response (Velagaleti et al., 2024, p. 2053). Such a framework highlights the connection between AI technologies and human emotions.

SEL Outcomes (Learning Policy Institute, 2024)

a 63% of studies assessed outcomes at post-test only (within 6 months of the end of intervention).
b 73% of studies assessed outcomes at post-test only (within 6 months of the end of intervention).
Note: * is p < 0.05
 

Evidence for SEL (Learning Policy Institute, 2023)

Although research (like the above meta-analyses and attached report) has proven social emotional learning to be an effective, there have been recent criticisms against this approach. Clark et al. argue that “SEL standards may undermine or erase the critically productive role that emotions have played in movements for social justice by ignoring racism, ableism, and other oppressions; privileging civility over productive conflict; and focusing on behaviors over emotions, especially when expressed by Black, Brown, dis/abled, and queer people” (Clark et al., 2022, p. 132). Some critics have also posited focusing on psychoanalyst Melanie Klein’s theory that is important to make space “for negative affect, aggression, and awareness of the body in the classroom, showing how working with and through these phenomena allows for creativity and learning” (Stearns, 2018, p. 8). Klein believed that “relational, creative, and educational opportunities get lost when the early childhood classroom aims insistently on ensuring the omnipresence of a specific version of happiness and of what it means to act and think” (Stearns, 218, p. 9).

Some like a 2021 Washington Examiner article, have even gone as far as suggesting that social emotional learning may be a “Trojan Horse” for teaching students about gender diversity and critical race theory and have accused it of being an indoctrination tool as part of a “woke agenda”. Supporters of SEL argue that these critics ignore the large amount of research proving the positive outcomes of a social emotional learning approach. Eleanor Bader states that “between 2021 and early 2023, at least 25 states saw bills introduced into their legislatures to remove social and emotional learning (SEL) from public school curricula” and “as of late 2023, eight states were considering bills to limit or ban SEL” (Bader, 2023). Many critics conflate critical race theory (CRT) and SEL, believing that SEL is used to deliver CRT’s ideology into school systems.

Social Emotional Learning and Mental Health Academy

There are many benefits of SEL programs as seen above. It is considered to play a very important role in a student’s learning and development. Upon examination of five evidence-based SEL programs, Pinar Aksoy and Frank Gresham state that (Aksoy & Gresham, 2024, p. 212):

The available experimental studies conducted on preschool-aged children found that intervention programs such as First Step to Success, I Can Problem Solve, Incredible Years (Dina Dinosaur Social Skills and Problem-Solving Curriculum), PATHS (Promoting Alternative Thinking Strategies), and Strong Start Pre-K were effective in improving various social-emotional learning skills. The results of these studies indicated that these programs substantially decreased problem behaviors (e.g., hyperactivity, anxiety, aggression, and/or conduct problems) in preschool children. Each of the programs contributed to decreasing children's social-emotional problems and fostering their social-emotional competence.

There are numerous proven benefits and advantages of using AI in the field of education. AI-driven educational tools provide an immersive and interactive learning experience which results in increased engagement in SEL. AI is able to analyze “large datasets of emotional responses and interactions, AI can generate valuable insights for educators to understand their students’ emotional needs, preferences, and trends over time, informing instructional decision-making (Pardos et al., 2021). Using adaptive algorithms, AI educational systems can personalize the approaches according to each student’s emotional needs and their respective learning styles. Tools using sentiment analysis, a subfield of natural language processing (NLP), offer real-time insights into a student’s emotional states, allowing educators to provide timely feedback, interventions, and support strategies to address emotional barriers pertaining to their learning (Altrabsheh et al., 2018).

SEAIL Model (Global EdTech)

AI-driven tools also promote inclusivity and equity in education by offering accessible and flexible support systems that can adapt to diverse learning styles, needs, and abilities. Thus, “by harnessing the potential of these technologies thoughtfully and ethically, educators can create inclusive and supportive learning environments that nurture students’ holistic development” (Sethi & Jain, 2024, p. 221). Below are the most popular forms of AI-driven tools being used in SEL.

AI Learning Platforms

Platforms like EdApp, Khan Academy, and Coursera are some of the AI-driven adaptive learning platforms that analyze students’ interactions, performance data, and emotional states, in order to dynamically adjust instructional content, pacing, and interventions to optimize learning outcomes (Blikstein, 2013). These platforms immensely help both students and teacher since they are able to give instant feedback on students’ assignments/assessments along with identify learning gaps that allows educators to focus on the learners’ weak areas. New learning materials can also be generated to support the existing teaching content. In addition, they can provide detailed analytics on a learner’s performance and makes it easy for teachers to track their progress. Accessibility features like text-to-speech and/or translations are also available to assist learners with such needs.

AI Chatbots/Virtual Assistants

AI-driven chatbots and virtual assistants can provide guidance and emotional support to students. These tools use “natural language processing (NLP) algorithms and machine learning techniques, these conversational agents can engage users in empathetic interactions, offer emotional validation, and deliver personalized interventions based on users’ emotional states and needs” (Sethi & Jain, 2024, p. 216). Chatbots and virtual assistants can be easily accessed and scaled on various platforms.

A meta-analysis conducted by Rong Wu and Zhonggen Yu resulted in the below findings: “AI chatbots could have a large effect on students' learning outcomes in terms of performance, motivation, interest, self-efficacy, perceived value of learning and anxiety. Additionally, this meta-analysis found that educational levels and intervention duration could moderate the effects of AI chatbots on learning outcomes. The results indicated that students in higher education seemed to benefit the most from AI chatbots. By contrast, primary and secondary school students assisted with AI chatbots could not have better learning outcomes than those without using AI chatbots. Regarding the intervention duration, short interventions were found to be more effective in enhancing students' learning outcomes, compared to long interventions” (Wu & Wu, 2023, p. 26).

AI Gamification

Another AI-driven tool that has been gaining traction is gamification, defined by Sethi and Jain as “the application of game design elements in non-game contexts” (Sethi & Jain, 2024 , p. 217). These games incorporate features like points, badges, and varying challenges to engage learners. Gamificiation can promote empathy in a learner by immersing them in storytelling experiences. This gives the learner/user the ability to explore different perspectives and complex social scenarios along with facilitating cooperative learning experiences that promote empathy and collaboratoin among the learners’ peers resulting in a supportive and inclusive learning environment (Sethi & Jain, 2024). Duolingo and Kahoot! are a few of the most popular AI gamification tools currently in use.

AI Robots/Toys

Emerging AI robotic technologies have the “potential to develop children’s social and emotional competencies and seek children’s improved participation in their natural learning environments” (Kewalramani et al., 2021, p. 367). Although such AI robots/toys cannot replace educators, these technologies should be combined with other natural interventions to help the student advance in their SEL goals. These robots have face and voice recognition with the ability to converse with learners. Research has shown “that the robot’s artificially intelligent behaviors (ability to talk, smile and move) and children’s interactions can produce a dramatic and emotive paradigm, which can help improve children’s empathic performances (Kewalramani et al., 2021, p. 356). The dynamic AI-capabilities of these robots enable them to sense and react to the learner’s responses resulting in playful interactions which provide a context for a playful and collaborative SEL experience, much like the below example in the video.

Challenges/Criticisms

AI has much potential along with many limitations. SEL features need to be incorporated into AI educational tools. As Walker explains that although "AI can automate many routine tasks, it cannot replicate essential human qualities, such as empathy, creativity, and critical thinking. Here’s why social emotional skills are needed to successfully navigate the complexities of an AI-driven society:

- Empathy and collaboration. As AI becomes more prevalent, human interaction and collaboration will remain vital. SEL helps young people develop empathy, understanding, and the ability to work effectively with others. These skills enable them to navigate diverse perspectives, build strong relationships, and collaborate in teams.

- Adaptability and resilience. In a world where technology is evolving rapidly, adaptability and resilience are key. SEL equips young people with the ability to adapt to change, bounce back from setbacks, and embrace continuous learning. These skills foster a growth mindset and enable individuals to thrive in dynamic and uncertain environments.

- Ethical decision-making. AI presents complex ethical challenges, such as privacy concerns, algorithmic bias, and job displacement. SEL helps young people develop ethical decision-making skills, enabling them to critically evaluate the impact of AI and make informed choices that prioritize human well-being, fairness, and social justice." (Walker, 2023)

The future of Artificial Intelligence in SEL is very bright, but some issues will continue to arise. Regarding this, Sethi and Jain state that “prioritizing ethical considerations and establishing clear guidelines are essential to ensure the responsible design, deployment, and evaluation of AI driven SEL technologies, safeguarding privacy, fairness, and transparency. Interdisciplinary collaborations between computer scientists, psychologists, educators, and ethicists can enrich the development process by integrating diverse perspectives and expertise. Longitudinal studies are crucial to assessing the long-term impact of AI interventions on students’ socio-emotional development, academic outcomes, and well-being. Culturally sensitive designs that account for individual differences and diverse contexts can promote inclusivity and effectiveness across diverse student populations” (Sethi & Jain, 2024, p. 221). If used properly (keeping in mind ethical/practical concerns), the prevalance of AI tools has the potential to strengthen the bond between a teacher and a student.

Future research should also “explore the long-term effects of SEL intervention programs on children’s social-emotional development” (Aksoy & Gresham, 2024, p. 213). Salas states “traditionally, research on learning outcomes has focused mostly on intellectual outcomes (cognitive), rather than supporting a more integral perspective that includes physical outcomes (psychomotor) and social-emotional outcomes (affective). Researchers should perhaps build upon the present study to further examine the interdependence among the attainment of different learning outcomes from social, psychological and technological perspectives” (Salas, 2020, p. 1822). It will always be a challenge for AI to understand the complexities of human emotions. The field of education will undoubtedly continue to be transformed by Artificial Intelligence. AI-driven tools should be used by learners under the supervision of caregivers or educators to protect learners from any negative effects on their creativity and or emotional growth.