DEVOPSdigest

As researchers focused on the technology industry, we commonly examine the perceptions that different roles have about emerging technology. We've been monitoring perceptions of Artificial intelligence (AI) for over a decade and early on we hypothesized that developers might be skeptical and have a difficult time trusting AI. We were wrong, developers commonly look to their community for help with coding techniques, troubleshooting issues and sharing insights.

Over the years we've seen increased openness from developers towards AI because of its ability to serve some of the same purposes that they have turned to sources like Stack Overflow, GitHub, Reddit and other online forums and communities in the past.

Developers are enthusiastic about the opportunities that AI presents for automating tasks, generating code snippets, and reducing development time. AI can even improve code quality by providing suggestions for code improvements and ensuring adherence to coding standards and best practices.

While AI has already made many strides in what it is capable of in the field of development, AI models rely heavily on data for training and validation. Outputs may only be as helpful as the underlying data; if it hasn't been trained on code examples for the task at hand or in the desired language, it won't be much assistance. AI also can lack transparency or have the potential to amplify biases present in training data, making it difficult for developers to understand or rely upon the recommendations coming out of AI.

"I use AI a lot and it helps me write better code. It's only useful if it is very familiar with the programming language."(Reddit)(link is external)

"The AI improves coding efficiency, many professional programmers have stated as much. Better AI would improve it even more." (Reddit)(link is external)

Some developers have expressed concerns about job displacement with some saying it will take away the need for coding as it exists today.

"AI will kill coding like compilers killed programmers."(Reddit)(link is external)

More likely AI will create a shift in developer's roles. New opportunities that didn't exist before will arise for developers who incorporate AI into their projects and focus on AI technology itself. Those developers with specialized skills focused on machine learning, natural language processing and data science are likely to be in even greater demand. AI also brings with it new prospects for developer collaboration with those in adjacent fields like data science, math, and ethics.

Corporations have been looking to low-code/no-code platforms for years in the attempt to enable those without extensive coding knowledge to do part of the job of a developer in theory. But as AI continues to mature, will companies rely more and more on the coding content generated by AI and downplay the need for human problem-solving skills and innovation?

Currently the common sentiment is that human review, assistance, and guidance from experienced developers is needed to evaluate code that is generated by AI.

"AI, even good AI is unreliable because it doesn't make something good, it makes something that looks good, regardless of if it works or not. Plus, code always needs debugging and updating. Scrutiny is something humans are particularly good at." (Reddit)(link is external)

Today, there are plenty of experienced coders who can evaluate and bring insight to how — and how not to — incorporate the code being generated by AI. However, as reliance on AI code generation perpetuates, will we see a decline in human coding capabilities?

Similarly, without developers honing their skills on the rigor of basic coding will they lose their understanding of the intricacies that allow them to create elegant and innovative solutions?

There is a risk of developers becoming overly dependent on AI, which could lead to a decline in new ideas and problem-solving abilities. To avoid this, it is important for developers and the companies that employ them to continue to view AI as a complement to existing skills, rather than a replacement and actively engage in continuous learning rather than seeing AI as a means to time and cost savings in the future.