Will AI kill DEV? Will prompt engineering replace coding?

Written by Daniel Eycken | June 24, 2024 @ 7:07 AM

By Daniel Eycken, partner-COO CIONET

This CIONET roundtable, held in the picturesque site De Abdijmolen in Leuven in partnership with CloudBees, brought together senior digital leaders, to tackle a provocative question: "Will AI kill development?". The event focused on whether prompt engineering might replace traditional coding: it gathered a diverse group of CIOs and software development managers to explore the implications of AI advancements for the future of software development.

Jean-Sébastien Ntigura (SWIFT): AI's potential and it's challenges

Jean-Sébastien Ntigura, Product Owner at Swift, provided an insightful perspective on AI's potential and its challenges. With over 20 years in application development, Jean-Sébastien highlighted how AI is already significantly accelerating software development cycles. He elaborated on AI's capabilities in domains such as auto-completion, debugging, and test automation. Swift already has a history of working with AI, particularly in fraud detection, showcasing AI’s potential to enhance its business objectives. Today, the organisation is exploring the broader use and impact of AI, prioritising the establishment of a clear and solid framework with stringent governance rules. This cautious approach stems from significant security challenges posed by AI. Jean-Sébastien emphasised the importance of maintaining air-gapped environments within Swift to protect sensitive data. He also highlighted the "black box" nature of AI, which complicates understanding the rationale behind certain “AI-decisions”, further underscoring the need for robust governance frameworks.

Stefan Simenon (Aegon): The Dual Nature of AI

Stefan Simenon, Director of Engineering at Aegon, echoed Jean-Sébastien's views on AI's dual nature as both an opportunity and a threat. At Aegon, AI is already used in multiple processes and it is being explored to further enhance continuous software delivery (CSD) processes. Stefan indicated several AI potentials but he also warned for increased application security vulnerabilities. Junior developers, must be well guided during implementation. Application security scanning will become more important. But he also stressed the importance of managing expectations within senior management to avoid over-estimations of AI's capabilities. He highlighted the necessity of clear policies and governance to mitigate risks associated with AI. The focus, he suggested, should be on making IT operations more efficient while uplifting the role of developers by freeing them from mundane tasks, thus allowing them to focus on higher-value activities.

Security and data protection

During the discussions, several key points emerged. Participants agreed that while AI holds promise for transforming software development, the journey is still in its early stages. Many organisations are currently conducting proof-of-concept (POC) projects to evaluate AI's potential. The primary objective for most is to enhance efficiency, particularly in areas such as test automation and security. These POCs have also shown an increased number of application security vulnerabilities. Security and data protection remain therefore significant concerns. Participants thus recommended establishing a control centre to monitor AI applications, similar to a Security Operations Centre (SOC).

Divide in embracing AI

A recurring theme was the generational divide in embracing AI. Younger developers are more

inclined to integrate AI into their workflows, seeing it as a tool that gives them a competitive edge. In contrast, older developers tend to be more sceptical. This generational gap highlights the need for comprehensive training programmes to ensure all developers can leverage AI effectively. The enthusiasm of younger developers for AI’s capabilities stands in stark contrast to the cautious approach of their more experienced counterparts, underscoring the need for targeted education and training to bridge this divide and ensure cohesive team dynamics.

AI’s potential to democratise technology

Another significant discussion point was AI’s potential to democratise technology. By lowering the barriers to entry, AI could enable a more diverse range of individuals to participate in software development. This democratisation could lead to increased innovation and a broader pool of talent. The fellowship at the table highlighted how AI could potentially revolutionise the development landscape by making advanced tools accessible to a wider audience, thus fostering a more inclusive environment.

Despite the optimism, there was also a sense of caution. Participants warned against becoming overly reliant on AI, which could lead to a loss of fundamental skills. They emphasised the need to maintain critical thinking and problem-solving abilities, ensuring developers can still perform essential tasks without AI. This is particularly important for resilience in crisis situations, where the ability to fall back on basic skills is crucial.

The economic implications of AI

The discussions also covered the economic implications of AI. While AI can streamline processes and reduce costs in the long term, the initial investment in AI technology and training can be substantial. Organisations need to balance these costs with the expected benefits, carefully planning their AI strategies to ensure a positive return on investment. Moreover, the evolving nature of AI technology means that continuous learning and adaptation are necessary to stay competitive.

Conclusion

In conclusion, the roundtable underscored that AI's impact on software development is profound but still unfolding. The discussions revealed a cautious optimism: while AI can undoubtedly enhance efficiency and capability, it also brings significant challenges that need to be managed carefully. The key takeaway is the importance of asking "why" when considering AI, ensuring that the use of AI aligns with organisational goals and values. As AI continues to evolve, it will be crucial for organisations to stay agile, continuously assess their strategies, and maintain robust governance to harness AI's potential effectively. The integration of AI into software development is not just about technological advancement but also about reshaping the role of developers and the structure of development processes to align with broader business objectives.

View full post