ai <- read_csv("AI_Student_Life_2026.csv")
## Rows: 100 Columns: 10
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (7): Gender, Education_Level, City, AI_Tool_Used, Purpose, Impact_on_Gra...
## dbl (3): Student_ID, Age, Daily_Usage_Hours
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.
ai_clean <- ai %>%
  select(-City) %>%
  mutate(
    Education_Level = as.factor(Education_Level),
    Gender = as.factor(Gender),
    AI_Tool_Used = as.factor(AI_Tool_Used),
    Purpose = as.factor(Purpose),
    Impact_on_Grades = as.factor(Impact_on_Grades),
    Satisfaction_Level = as.factor(Satisfaction_Level)
  )

Learning With Machines: How Generative AI Is Changing Student Learning

Introduction

Generative AI tools such as ChatGPT are rapidly transforming modern education. Students increasingly rely on AI for writing, coding, research assistance, summarising information and study support. What was once considered experimental technology has quickly become embedded within everyday academic workflows.

At the same time, the growing dependence on AI raises important questions for universities. Does AI genuinely improve learning outcomes, or does it mainly improve speed and convenience? Are all students using AI equally? And how might universities adapt as AI becomes a permanent part of student learning?

This story explores patterns of AI usage among students through a series of interactive visualisations examining usage frequency, educational differences, academic outcomes and student satisfaction.

Chart 1: AI has become part of everyday student study routines

Many students are no longer using AI occasionally. Instead, AI tools are becoming part of daily academic behaviour. This chart highlights how frequently students engage with generative AI during study activities.

p1 <- ai_clean %>%
  mutate(Usage_Group = cut(
    Daily_Usage_Hours,
    breaks = c(0, 1, 2, 3, 4, 5, Inf),
    labels = c("Up to 1 hour", "1–2 hours", "2–3 hours", "3–4 hours", "4–5 hours", "More than 5 hours"),
    include.lowest = TRUE
  )) %>%
  count(Usage_Group) %>%
  ggplot(aes(
    x = Usage_Group,
    y = n,
    text = paste(
      "Usage:", Usage_Group,
      "<br>Students:", n
    )
  )) +
  geom_col(fill = charcoal) +
  labs(
    title = "AI tools are now part of daily study behaviour",
    x = "Daily AI usage",
    y = "Number of students"
  ) +
  theme(
    axis.text.x = element_text(angle = 30, hjust = 1)
  )

ggplotly(p1, tooltip = "text")

Chart 2: Postgraduate students rely on AI more heavily

AI usage is not evenly distributed across all students. Differences emerge across education levels and gender, suggesting that academic demands and learning styles may shape how students interact with AI technologies.

p2 <- ai_clean %>%
  group_by(Education_Level, Gender) %>%
  summarise(avg_hours = mean(Daily_Usage_Hours), .groups = "drop") %>%
  ggplot(aes(
    x = Education_Level,
    y = avg_hours,
    fill = Gender,
    text = paste(
      "Education level:", Education_Level,
      "<br>Gender:", Gender,
      "<br>Average daily AI usage:", round(avg_hours, 2), "hours"
    )
  )) +
  geom_col(position = "dodge") +
  scale_fill_manual(values = c(indigo, tomato)) +
  labs(
    title = "Average AI usage differs across education levels",
    x = "Education level",
    y = "Average daily AI usage (hours)",
    fill = "Gender"
  )

ggplotly(p2, tooltip = "text")

Chart 3: Students use AI for far more than essay writing

Public discussions about AI in education often focus heavily on cheating and essay generation. However, students use AI for a much broader range of purposes including coding assistance, research support, brainstorming and summarising information.

p3 <- ai_clean %>%
  count(Education_Level, Purpose) %>%
  group_by(Education_Level) %>%
  mutate(percent = n / sum(n) * 100) %>%
  ggplot(aes(
    x = Education_Level,
    y = percent,
    fill = Purpose,
    text = paste(
      "Education level:", Education_Level,
      "<br>Purpose:", Purpose,
      "<br>Percentage:", round(percent, 1), "%"
    )
  )) +
  geom_col() +
  scale_fill_brewer(palette = "Set2") +
  labs(
    title = "Students are integrating AI into many forms of academic work",
    x = "Education level",
    y = "Share of students (%)",
    fill = "Purpose"
  )

ggplotly(p3, tooltip = "text")

Chart 4: More AI use does not always produce better academic outcomes

While many students believe AI improves productivity and efficiency, the relationship between AI usage and academic outcomes appears more complicated. Increased AI use does not necessarily guarantee stronger learning performance.

p4 <- ai_clean %>%
  ggplot(aes(
    x = Impact_on_Grades,
    y = Daily_Usage_Hours,
    fill = Impact_on_Grades,
    text = paste(
      "Impact on grades:", Impact_on_Grades,
      "<br>Daily AI usage:", Daily_Usage_Hours, "hours"
    )
  )) +
  geom_boxplot(alpha = 0.8) +
  scale_fill_manual(values = c(indigo, jungle, tomato)) +
  labs(
    title = "Heavy AI usage does not automatically lead to better grades",
    x = "Reported impact on grades",
    y = "Daily AI usage (hours)"
  ) +
  theme(
    legend.position = "none"
  )

ggplotly(p4, tooltip = "text")

Chart 5: Student satisfaction with AI varies across platforms

Not all AI tools produce the same student experiences. Satisfaction levels vary across platforms, suggesting that usability, trust and perceived usefulness strongly shape how students evaluate AI technologies.

p5 <- ai_clean %>%
  count(AI_Tool_Used, Satisfaction_Level) %>%
  group_by(AI_Tool_Used) %>%
  mutate(percent = n / sum(n) * 100) %>%
  ggplot(aes(
    x = AI_Tool_Used,
    y = percent,
    fill = Satisfaction_Level,
    text = paste(
      "AI tool:", AI_Tool_Used,
      "<br>Satisfaction:", Satisfaction_Level,
      "<br>Percentage:", round(percent, 1), "%"
    )
  )) +
  geom_col() +
  scale_fill_brewer(palette = "Pastel1") +
  labs(
    title = "Student satisfaction differs across AI platforms",
    x = "AI tool used",
    y = "Share of students (%)",
    fill = "Satisfaction level"
  ) +
  theme(
    axis.text.x = element_text(angle = 30, hjust = 1)
  )

ggplotly(p5, tooltip = "text")

Conclusion

Generative AI is no longer a future issue for universities — it is already deeply integrated into student learning practices. The findings suggest that students are not simply experimenting with AI occasionally, but increasingly embedding AI tools into writing, coding, research and study support activities.

However, the relationship between AI use and learning outcomes remains uncertain. While many students report increased efficiency and convenience, higher AI usage does not consistently correspond with stronger academic performance.

For universities, the challenge is no longer whether students will use AI. Instead, the challenge is preparing students to use AI critically, ethically and effectively in a future where human learning and machine assistance are becoming increasingly interconnected.

Acknowledgements

ChatGPT was used to assist with coding support, structural planning, debugging and narrative refinement during the development of this assignment. All analysis, interpretation, verification and final editing were completed by the author.

References

Dataset Author. (2026). AI Student Life Dataset 2026 [Data set]. Kaggle. https://www.kaggle.com/datasets/guriya79/how-ai-is-changing-student-life

Code

All code used to preprocess the data and create the visualisations is included within this RMarkdown document.