The software development landscape is undergoing a significant shift as generative AI and AI coding assistants become more prevalent. A recent survey of over 300 business leaders reveals that while only 12% believe generative AI has fundamentally changed software development today, 38% predict it will substantially impact the software development lifecycle (SDLC) across most organizations within the next one to three years. The adoption of generative AI in software development is widespread, with 94% of respondents using it in some capacity.
However, the level of integration varies, with nearly one-third still conducting small pilots or individual-employee adoption rather than team-wide integration. Despite this, 46% of leaders say generative AI is already meeting expectations, and 33% say it exceeds or greatly exceeds expectations. Generative AI is being used beyond code generation, with 82% of respondents using it in at least two phases of the SDLC.
Common use cases include designing and prototyping new features, streamlining requirement development, fast-tracking testing, and improving bug detection and code quality. Looking to the future, 49% of leaders believe advanced AI tools, such as assistants and agents, will lead to efficiency gains or cost savings, while another 20% believe they will lead to improved throughput or faster time to market. However, the rapid adoption of AI in software development has also led to challenges in maintaining security.
Generative AI adoption in software development
Developers are on track to download more than 6.6 trillion software components in 2024, and the mean time to remediate vulnerabilities has grown significantly over the past seven years, from about 25 days in 2017 to more than 300 days in 2024. Security researchers have warned that AI code generation could result in more vulnerabilities and novel attacks.
Developers also have growing concerns over the potential for AI assistants to suggest or propagate vulnerable code. While 56% of developers expect AI assistants to provide usable code, only 23% expect the code to be secure. The reliance on AI to complete simple programming projects could also reduce the need for new or entry-level developers who typically tackle simpler coding tasks, removing a training path.
This poses potential risks to younger generations of developers. To address these challenges, companies behind AI assistants need to create training datasets that contain secure code suggestions and put in place guardrails to protect against vulnerable and malicious code generation. Companies will also have to deploy automated software security tools to check the work of any coding assistant.
Despite these challenges, the future of software development looks promising with the ongoing advancements and deeper integration of generative AI. The security of software and applications could eventually become much stronger as a result of these developments.
Cameron is a highly regarded contributor in the rapidly evolving fields of artificial intelligence (AI) and machine learning. His articles delve into the theoretical underpinnings of AI, the practical applications of machine learning across industries, ethical considerations of autonomous systems, and the societal impacts of these disruptive technologies.
