Necessary cookies are absolutely essential for the website to function properly. These cookies ensure basic functionalities and security features of the website, anonymously.

Discussing the impact of Generative AI on Student Education
Generative AI has rapidly advanced to the point where it can generate entire sections of code, offering convenience to data science students and professionals alike. With tools such as GitHub Copilot, ChatGPT, and other AI-driven platforms, coding has become more accessible than ever before. However, this convenience raises significant questions in education: Are students still acquiring the coding and problem-solving skills necessary to succeed? Could reliance on AI tools erode originality and stifle the development of deeper technical expertise?
In this blog, we’ll explore these concerns, consider practical solutions, and reflect on how educators can continue to ensure students build the foundational skills necessary for success in the field of data science—despite the increasing presence of generative AI.
The Challenge of Over-Reliance on AI
Generative AI presents a dilemma. It allows students to produce working code more quickly and with fewer errors, but it also creates the temptation to bypass the learning process. Just as students in mathematics may rely too heavily on calculators, data science students may find themselves leaning on AI-generated code without fully understanding the reasoning or structure behind it.
The risk here is clear: if students consistently use AI to write code for them, they may miss out on developing the core coding and problem-solving skills they’ll need in professional environments. A strong foundation in coding and algorithmic thinking is essential, especially in a field like data science, where real-world problems are often novel, complex, nuanced, and require creative solutions.
Encouraging Reverse Engineering of AI-Generated Code
One of the most effective ways to circumvent this issue is by encouraging students to reverse engineer AI-generated code. The idea is simple: instead of treating AI output as a finished product, students should be required to dissect and explain the code. This process forces students to engage with the logic, syntax, and structure of the code, ensuring they understand not just what the code does, but why it does it.
For example, consider a data science student who uses AI to generate code for a decision tree model. Instead of moving on to the next task, the student could be asked to explain key parameters such as max_depth and min_samples_split. Why were these particular values chosen? How do they impact the model’s accuracy and efficiency? This deeper dive into the model’s mechanics forces students to engage with machine learning principles, allowing them to develop a more meaningful understanding of how their model behaves in different scenarios.
This reverse-engineering method is analogous to how students in chemistry may analyse reaction mechanisms rather than simply memorising equations. By dissecting the output, students become active participants in their learning rather than passive users of AI.
Designing Projects That Demand Creative Problem Solving
Another strategy to maintain coding proficiency is to design projects that require students to tweak or refine AI-generated code. Generative AI often produces code that, while functional, might not be optimal. Encouraging students to improve upon this code teaches them how to recognise inefficiencies and troubleshoot issues—skills that are essential for real-world data science.
For instance, in a machine learning class, students could be given AI-generated code for a basic neural network. The task could then be to optimise this network by adjusting hyperparameters or implementing more advanced features such as dropout for regularisation. In this scenario, students are compelled to think critically about how to improve performance, which fosters deeper learning and critical thinking.
This mirrors practices in many scientific fields. In biology, students might be asked to modify experimental conditions to achieve optimal results. Similarly, data science students can be challenged to refine AI-generated code to better fit the problem at hand.
Prioritising Coding from Scratch
Despite the convenience of AI, it is crucial that data science students are still given opportunities to code from scratch. Much like how medical students must practise procedures without the aid of advanced technology, aspiring data scientists must learn to build algorithms without the assistance of AI-generated templates. This approach ensures they develop an understanding of foundational concepts such as data structures, algorithms, and computational complexity, all of which are critical in the world of data science.
A common example might be asking students to implement a sorting algorithm such as quicksort or mergesort without relying on pre-built libraries. While these algorithms are rarely implemented manually in a professional setting, understanding their mechanics provides crucial insights into computational efficiency and algorithm design. Similarly, coding machine learning models from scratch—like logistic regression or k-nearest neighbours—gives students the chance to understand what’s happening behind the scenes, especially in terms of optimisation and gradient descent.
Focusing on Reasoning, Not Just the End Product
One of the most effective ways for educators to combat the over-reliance on AI tools is to shift the focus away from the final product and towards the reasoning and process behind it. In other words, it’s not enough for students to submit functional code; they must be able to explain their approach, their decisions, and the rationale behind the code.
Assessments that focus on process over product push students to engage more deeply with the material. For example, students could be asked to submit a written explanation alongside their code, detailing their decision-making process, the problems they encountered, and how they solved them. Alternatively, educators could conduct one-on-one discussions where students are asked to defend their code and reasoning. This approach prioritises critical thinking, ensuring that students are developing the problem-solving skills necessary for more complex tasks, not just relying on AI for a quick solution.
This is akin to how problem-solving is often evaluated in mathematics—students are graded not only on whether they got the correct answer, but also on how they arrived at it. In data science, this ensures students understand why a solution works, rather than simply accepting that it does.
Ethical Considerations in AI Usage
Ethical considerations also play a significant role in helping students develop a responsible relationship with generative AI. As future data scientists, students will face ethical challenges related to data privacy, bias in algorithms, and the broader societal impact of AI technologies. Encouraging discussions about these issues helps students understand the implications of over-reliance on AI and reinforces the importance of developing a strong technical foundation.
In the same way that engineers must consider the ethical implications of automation, data scientists must be mindful of how AI influences their work. For example, students can explore case studies where AI tools have been misused or resulted in unintended consequences, discussing how a lack of understanding of the underlying technology led to negative outcomes. This approach ensures students think critically about their use of AI, both in their studies and future careers.
Encouraging Creativity and Innovation
Ultimately, creativity is what will set apart the next generation of data scientists. While AI tools can assist with coding, they won’t replace the need for innovative thinking. Data science often requires novel approaches to problem-solving, and educators can support this by assigning projects that demand creative solutions.
For example, students could be tasked with designing an original machine learning model for a unique dataset, or developing a new way to visualise complex data. These kinds of open-ended challenges encourage students to go beyond pre-existing templates, fostering an environment where generative AI becomes a tool for inspiration, rather than a substitute for creativity.
Much like how architects may use software to generate standardised plans but rely on their creative vision to produce something unique, data scientists should be encouraged to innovate, using AI as a supportive tool rather than the entire solution.
Tools for checking validity of student work
There are also tools being developed to detect the use of generative AI in academic writing, similar to plagiarism detection software. However, these tools are still in their infancy and currently have a tendency to produce false positives, flagging genuine student work as AI-generated. While these tools may evolve over time, it reinforces the importance of focusing on the learning environment and the opportunities we provide, rather than solely relying on detection mechanisms. At the end of the day, creating a setting where students feel motivated to learn and apply their skills is the most effective way to prepare them for real-world challenges.
Conclusion
Generative AI is undoubtedly moulding the way coding is taught and practised, especially in the field of computing sciences. However, as educators and students, we must ensure that it remains a tool for learning and not a shortcut that bypasses the development of essential skills. By focusing on the process of coding, encouraging reverse engineering, designing problem-solving projects, and ensuring students still have opportunities to code from scratch, we can help students build a strong foundation that will serve them well in their future careers. Additionally, by fostering creativity and encouraging ethical discussions, we ensure that the next generation of data scientists is not only technically proficient but also innovative and responsible in their use of AI.
As educators, we must also acknowledge that there’s only so much we can control in terms of how students choose to complete their academic work. With the availability of AI tools, some students may be tempted to use shortcuts or rely heavily on generative AI, regardless of our best efforts. All we can do is place students in the best environment possible to encourage authentic learning and development. This involves providing them with engaging, challenging tasks and fostering a culture that values deep understanding over quick fixes.
Generative AI is powerful, but the human mind remains unmatched in its capacity for creativity, problem-solving, and ethical decision-making.
Join the IoA and prepare for the future of business
Sign up now to access the benefits of IoA Membership including 1400+ hours of training and resources to help you develop your data skills and knowledge. There are two ways to join:
Corporate Partnership
Get recognised as a company that works with data ethically and for investing in your team
Click here to joinIndividual Membership
Stand out for your commitment to professional development and achieve the highest levels
Click here to join