You think AI will render coding redundant? Not so fast


(Original article is featured here: https://www.straitstimes.com/opinion/you-think-ai-will-render-coding-redundant-not-so-fast
(Straits Times, 19 March 2024)

Written By:

Professor Hahn Jungpil
Vice Dean, Communications
Director, NUS Fintech Lab
Deputy Director (AI Governance), AI Singapore
Deputy Director, Centre for Technology, Robotics, Artificial Intelligence & the Law

In an era where artificial intelligence (AI) is no longer a specialized topic of academic interest but has now become central to our technological landscape, Nvidia’s CEO Jensen Huang’s recent comments at the World Government Summit has sparked a significant debate with his assertion that learning programming, or more broadly, computer science may no longer be necessary. This proposition, naturally, raises eyebrows and stirs unease among students and parents who have invested heavily in coding education, viewing it as a crucial skillset for the future. Yet, this discourse opens a vital avenue for nuanced discussion about the evolving role of education in the age of artificial intelligence. While Huang’s statement carries a kernel of truth regarding the utility of AI in programming, it simultaneously oversimplifies the multifaceted nature of computing education and its broader implications on cognitive development and problem-solving skills.

At the heart of the controversy is the assertion that advancements in AI will enable individuals without formal training in programming or computer science to generate working programs. There is definitely merit to this argument; indeed, AI’s capabilities to understand and generate code based on user inputs are already quite impressive, and these capabilities have the potential to democratize access to programming by lowering barriers for domain specialists without formal programming training. This shift could potentially alter the landscape of innovation development, making it more accessible to a broader audience. However, this vision, while compelling, overlooks the intricate complexity of creating robust solutions to complex problems.

AI-generated programs, while already quite impressive, are inherently limited by the specificity and scope of the prompts they receive from users. These tools may excel at addressing well-defined, narrow tasks but falter when faced with the necessity for broad, flexible thinking and anticipation of myriad possible scenarios. AI-generated programs may work flawlessly for the scenarios envisioned in the users’ prompts but remain woefully inadequate for unspecified and thus unanticipated scenarios which inevitably leads to significant technical debt if these programs are deployed. This gap underscores a fundamental misunderstanding of the essence of programming as not merely code generation but the cultivation of a deep, nuanced understanding of the complex problem space and the conceptual rigor required to navigate them.

Equating computing education solely with learning to write code is a reductive view that overlooks the computing discipline’s core value: computational thinking. Computational thinking is a cognitive framework that extends far beyond the syntax of programming languages; it embodies a universal problem-solving methodology that is applicable across disciplines. Computational thinking is about structuring and processing information and formulating solutions to complex problems in a systematic, logical manner. It involves decomposing complex problems into manageable parts, abstracting general principles from specific instances, and designing generalized algorithms to solve a wide array of problems. This intellectual rigor in approach to problem-solving is what computing education truly aims to instill. It’s about training minds to systematically approach problems and devise generalized problem solving strategies that are both effective and efficient. Such skills are invaluable, not only in the realm of technology but in any field that requires innovation through critical thinking.

Recognizing the limitations of AI-generated programs and the broader applicability and value of computational thinking leads us to rethink educational strategy to embrace the complementary roles of AI and human intellectual capabilities. In a world where technology increasingly intersects with every field of study, whether it be healthcare, finance, or the arts, the combination of computing skills with domain-specific knowledge becomes a powerful asset.  An interdisciplinary approach, such as a double major or a significant commitment to both computing and a non-computing field, offers a promising path for students. This educational strategy will enable them to harness the problem-solving methodologies of computational thinking while deeply understanding the nuances of a specific domain.

Such synergies between computational capabilities and domain expertise are crucial for devising complete, robust solutions to complex problems. They allow for a more nuanced understanding of the problem space, fostering innovation and creativity. This approach not only makes graduates more versatile and adaptable but also better equipped to leverage AI tools effectively, using them as partners in the problem-solving process rather than as replacements for human intellect.

Jensen Huang’s comments, while stirring controversy, serve as a catalyst for a much-needed conversation about the future of education in an AI-enabled world. The advent of AI in programming does not herald the obsolescence of computing education; rather, it highlights the need for a curriculum that balances technical skills with the development of computational thinking capabilities. As AI continues to reshape the professional landscape, the most successful individuals will be those who can seamlessly integrate computational thinking with domain-specific knowledge. By pursuing an interdisciplinary education, students can position themselves at the forefront of this transformation, equipped not just to adapt to the changes AI brings but to lead them. In doing so, they will prove that the value of education lies not in the rote learning of specific skills, but in the cultivation of a versatile, adaptive intellect capable of navigating the complexities of a rapidly evolving world.