AI-Assisted Learning for Improved Student Engagement: Lessons from the Three SSSs

1. Introduction

The integration of Artificial Intelligence (AI) in education has profoundly reshaped the delivery and experience of learning across disciplines, particularly within economics. Advances in AI, exemplified by interactive platforms such as ChatGPT and analytical software such as RStudio, have catalysed new modes of student engagement by facilitating immediate and adaptive feedback, enabling personalised learning pathways, and fostering active cognitive participation (Luckin and Holmes, 2016; Holmes et al., 2019). This transformative approach aligns closely with contemporary pedagogical models that underscore the importance of active learning, scaffolding, and constructivist principles, which collectively aim to move learners from passive information recipients to active participants and co-creators of knowledge (Biggs & Tang, 2022).

Within higher education institutions such as De Montfort University, pedagogical innovations have been actively pursued to optimise student engagement, retention, and overall learning outcomes. One notable advancement in this area is the development and implementation of the Three SSSs pedagogy (“See it, Say it, Sorted”), designed specifically to enhance student interaction, articulate learning challenges clearly, and expedite resolution through structured instructor and peer intervention. This pedagogical model inherently fosters deeper cognitive processing by compelling learners to actively identify, articulate, and resolve educational hurdles through systematic steps.

Building on the foundational principles of the Three SSSs model, this case study critically investigates the integration and efficacy of AI-assisted learning within this structured pedagogical framework, particularly focusing on Economics modules such as econometrics and macroeconomic analysis. Utilising empirical data, including real-time prompts, screenshots from student interactions, and reflective observations from teaching practice, the study provides a nuanced exploration of how iterative, AI-supported learning activities influence student engagement, motivation, and learning outcomes. The analysis addresses the evolving dynamics between student-led inquiry and AI-generated assistance, demonstrating how iterative cycles of discovery, articulation, and solution underpin effective learning interactions and critical thinking development.

2. Model Framework: Integrating the Three SSSs into AI-Assisted Learning

The Three SSSs ("See it, Say it, Sorted") pedagogy is systematically embedded into AI-Assisted Learning to maximise student engagement and critical thinking. This integration leverages AI tools to foster independent exploration while providing structured tutor intervention to resolve learning barriers.

Phase 1: See it (Identifying and Initial Engagement)

In this initial phase, students proactively engage AI chatbots by inputting specific, guided prompts related to their learning objectives. The chatbot generates immediate responses such as econometric codes or explanations relevant to RStudio statistical analysis. Students then actively examine these AI-generated responses, critically assessing their initial accuracy and relevance. This step facilitates early identification of conceptual misunderstandings or coding errors and prompts deeper cognitive engagement, aligning with constructivist learning theories (Fosnot & Perry, 1996).

Phase 2: Say it (Articulating Challenges and Errors)

Upon executing the generated RStudio code, students frequently encounter operational errors or discrepancies. Following the Three SSSs model, students are encouraged to immediately articulate these encountered challenges, explicitly stating the nature and context of errors. By voicing their concerns or confusion, students take active ownership of their learning challenges, thereby fostering an essential metacognitive skill and supporting the scaffolding principles identified by Wood, Bruner & Ross (1976).

Phase 3: Sorted (Collaborative Resolution and Feedback)

In response to articulated challenges, the instructor provides targeted feedback to students by interpreting and clarifying AI-generated code or theoretical explanations. The instructor guides students through the logic of correcting their prompts and adjusting coding syntax or analytical approaches, enhancing students' problem-solving capabilities. This step closes the learning loop by empowering students to iteratively refine their interactions with AI chatbots, thus reinforcing knowledge application and promoting deeper cognitive processing.

Figure 1: Model of AI-Assisted Learning Integrated with the Three SSSs Pedagogy

3. Discussion of Case Examples

To evaluate AI integration into the Three SSSs pedagogy, we explored a use case from an econometrics module, where students used RStudio to analyse macroeconomic data. Students initiated prompts to ChatGPT for package installation, code generation, data analysis, and visualisation. The Three SSSs model became particularly evident as students iteratively engaged in problem-solving by refining prompts, interpreting responses, and applying learned insights to their coding tasks.

Prompt Example 1: Initial Query

"Please provide me with all relevant packages to be installed on my RStudio for econometric analysis on the topic: Examining Debt Overhang and Macroeconomic Adjustments Impacts on Economic Growth in ECOWAS Member Countries."

AI Response

install.packages(c(
    "tidyverse", "readx1", "writexl", "lubridate", "ggplot2",
"dplyr", "janitor"
))

This phase represents the "See it" stage, where students identify the analytical resources, they need. Following the AI's suggestion, students attempted to load libraries and encountered a new set of errors due to syntax or undefined objects in their dataset.

Prompt Example 2: Code Execution Error

Error: object 'GDP_Growth' not found

Here, the AI's suggestion failed due to incorrect variable references in the dataset. This prompted learners to rephrase the query and clarify their dataset structure, reflecting the "Say it" phase, where students voice their confusion or misalignment between the dataset and the AI-provided code.

Prompt Revision

"Please generate R Studio code to clean column names and rename variables like GDP_Growth Rate, Inflation Rate, Public Debt, etc."

Corrected Code

nigeria_data <- nigeria_data %>%
  dplyr::rename(
    Year = Year,
    GDP_Growth = GDP.Growth.Rate,
    Inflation = Inflation.Rate,
    Public_Debt = Public.Debt
  )

This adjustment demonstrates deepened student understanding, increased prompt precision, and improved interaction with the AI. It also illustrates the essential role of iterative feedback in supporting learning. The back-and-forth between student and chatbot, supplemented by instructor clarification, exemplifies the pedagogy’s “Say it” and “Sorted” loop.

Prompt Example 3: Visualisation and Output

"Plot GDP Growth over time using ggplot2."

Error: could not find function "ggplot2"

Students responded with:

"Why is ggplot not working?"

AI Clarification

library(ggplot2)   #Make sure ggplot2 is loaded before use

AI-generated output

ggplot(nigeria_data, aes(x = Year, y = GDP_Growth)) +
  geom_line(color "blue", size = 1) +
  labs(title = "GDP Growth Rate Over Time", x = "Year",
    y = "GDP Growth (%)")
+
theme_minimal ()

At this point, the "Sorted" stage is achieved as the AI returns a working code block that enables students to visualise trends and explore real economic data. This visualisation allows students to meaningfully interpret macroeconomic patterns in ECOWAS countries.

The value of this iterative AI-student interaction lies in its power to mimic the scientific process: hypothesis formulation, testing, error discovery, and resolution. Students develop higher-order thinking by critically analysing AI responses, testing them in RStudio, recognising errors, and refining their queries. These feedback loops not only deepen engagement but also promote independent learning and problem-solving resilience—skills central to modern economics education.

Prompt Example 4: Descriptive Statistics with Multiple Packages

"Please generate the R Studio code to perform descriptive statistics for my dataset."

AI Code

install.packages(c("readxl", "psych", "moments"))
library(readxl)
library(psych)
library(moments)

nigeria_data <- read_excel("Nigeria.xlsx")
names(nigeria_data) <- make.names(names(nigeria_data))

summary.stats<- psych: describe(nigeria_data[,2:7])

This enabled students to extract comprehensive descriptive statistics, including skewness and kurtosis, which are often challenging for beginners.

Prompt Example 5: Regression Analysis

"Generate a multiple linear regression model examining GDP growth as a function of inflation and public debt."

AI Code

model <- lm(GDP_Growth ~ Inflation + Public_Debt,
  data = nigeria_data)
summary(model)

This interaction not only introduced students to econometric modelling but also facilitated discussions about statistical assumptions, multicollinearity, and model interpretation. These queries reflect the evolving improvement of students’ interactions with AI, gradually transitioning from basic commands to nuanced diagnostic troubleshooting and methodological reasoning. Each engagement cycle reinforces the Three SSSs model, promoting deeper cognitive engagement and metacognitive awareness through AI-supported discovery, expression, and feedback.

4. Reflections on Student Learning

Integration of the Three SSSs framework into AI-assisted learning has yielded substantial pedagogical benefits, particularly in fostering active student engagement, independent learning, and critical thinking. The immediacy of AI-generated feedback allows students to iteratively test hypotheses, receive corrective guidance, and visualise outcomes in real-time. This immediacy transforms the traditional learning dynamic, enabling students to take greater ownership of their learning processes.

Student reflections reveal heightened levels of motivation and confidence in using AI for complex tasks such as data cleaning, visualisation, and econometric modelling. The iterative cycles of prompt refinement and code troubleshooting contribute to a deeper understanding of both statistical programming and the economic concepts under study. This aligns with the findings of Zawacki-Richter et al. (2019), who argue that AI Chatbots can enhance learner autonomy and cognitive engagement when used within supportive pedagogical structures. Additionally, AI's ability to adapt responses to different learning styles aligns with evidence by Chen et al. (2020) who found that AI-based tools improve comprehension, especially in technical fields like economics.

Moreover, the AI-assisted learning process developed digital literacy, particularly in crafting precise prompts and interpreting output. While students initially struggled with vague queries or misunderstood outputs, these challenges became valuable learning experiences. As Holmes et al. (2019) contend, the educational value of AI lies not in delivering perfect answers, but in catalysing inquiry, experimentation, and reflective dialogue.

Importantly, the “Say it” and “Sorted” phases of the Three SSSs encouraged peer dialogue and tutor intervention, creating a community of inquiry (Garrison et al., 1999). This collaborative engagement ensured that students were not working in isolation but were supported through their learning journey, which reinforces socio-constructivist learning principles.

5. Conclusion

This case study has demonstrated the powerful synergies between Artificial Intelligence (AI)-driven platforms and structured pedagogical models by examining how the Three SSSs framework "See it, Say it, Sorted", can be purposefully integrated into economics education to enhance student engagement and autonomous learning. Grounded in constructivist and scaffolding learning theories, the Three SSSs pedagogy provides a systematic instructional model that encourages students to explore, articulate, and resolve their learning challenges iteratively. When combined with AI tools like ChatGPT and the RStudio analytical environment, this pedagogy facilitates deeper inquiry, experimentation, and reflective learning.

Through empirical illustrations and real-world teaching observations, we have shown that AI-assisted learning is not merely a mechanism for automation or passive information retrieval. Rather, it functions as a dynamic cognitive scaffold that supports hypothesis testing, adaptive problem-solving, and just-in-time learning. Our case examples from an econometrics module underscore how the prompt-feedback loop, reinforced by the Three SSSs, drives students to engage critically with both data and AI-generated responses, enhancing their capacity for digital fluency, statistical analysis, and domain-specific reasoning.

Importantly, the findings suggest that while AI can amplify student learning, its true pedagogical value is contingent upon structured frameworks that foster metacognition, critical inquiry, and social interaction. The Three SSSs acts as a pedagogical glue, ensuring that the use of AI does not devolve into superficial engagement but instead becomes a reflective and dialogic process that supports learning autonomy and resilience.

Nevertheless, successful integration of AI into higher education requires the development of prompt engineering skills, ethical data practices, and instructional literacy to help students interpret AI-generated outputs effectively. Institutions must also invest in professional development for educators to ensure that they can guide and moderate these evolving learning experiences.

References

Biggs, J. and Tang, C., 2022. Teaching for quality learning at university: What the student does. 5th ed. Maidenhead: Open University Press.

Chen, L., Chen, P. and Lin, Z., 2020. Artificial intelligence in education: A review. IEEE Access, 8, pp. 75264–75278. https://doi.org/10.1109/ACCESS.2020.2988510

Fosnot, C.T. and Perry, R.S., 1996. Constructivism: A psychological theory of learning. In: C.T. Fosnot, ed. Constructivism: Theory, perspectives, and practice. New York: Teachers College Press, pp. 8–33.

Garrison, D.R., Anderson, T. and Archer, W., 1999. Critical inquiry in a text-based environment: Computer conferencing in higher education. The internet and higher education, 2(2-3), pp. 87-105. https://doi.org/10.1016/S1096-7516(00)00016-6

Holmes, W., Bialik, M. and Fadel, C., 2019. Artificial intelligence in education promises and implications for teaching and learning. Center for Curriculum Redesign.

Luckin, R. and Holmes, W., 2016. Intelligence unleashed: An argument for AI in education.

Wood, D., Bruner, J.S. and Ross, G., 1976. The role of tutoring in problem solving. Journal of Child Psychology and Psychiatry, 17(2), pp. 89–100. https://doi.org/10.1111/j.1469-7610.1976.tb00381.x

Zawacki-Richter, O., Marín, V.I., Bond, M. and Gouverneur, F., 2019. Systematic review of research on artificial intelligence applications in higher education – where are the educators? International Journal of Educational Technology in Higher Education, 16(1), p. 39. https://doi.org/10.1186/s41239-019-0171-0