How AI will reshape software development

The impact artificial intelligence will have on the digital universe should not be understated. Not since the development of the electrical grid have humans faced such a radical shift in their relationship to the natural world.

And while many leading experts tout the ways AI will alter key functions that we all take for granted — everything from transportation to buying new shoes — it appears the first realm to face a new AI-driven operating paradigm is the one that created AI in the first place: software development.

Expanding intelligence

According to IDC, the worldwide AI market is expected to top $341.8 billion in 2021 and then blow past $500 billion by 2024 for an annual growth rate of 18.8%. It should be noted, however, that fully 88% of this market will be in the form of software, about half from applications.

How will this change the software market? For one thing, it will alter the way code is written, updated, and released. With AI empowered to make changes to itself, the focus of highly trained, highly paid software engineers will shift from the dull drudgery of writing and rewriting minute functions to more creative, strategic-level operations that produce greater value and drive core business operations. At the same time, DevOps will become increasingly automated and responsive to users with the power to define their objectives and have AI convert them into code.

This could alter the way software is bought and sold, with individual users getting the updates they need from AI, rather than having to shell out more dollars for a generic new release from the developer.

This does not mean the end of commercial software development, however. As tech writer Smith Johnes notes, AI will help companies become leaner and meaner in a number of ways. Starting with more generic AI benefits, like better market analysis and cost-planning, firms can also look forward to AI-based programming assistants, automated testing and compiling, and streamlined bug fixing — all of which promises to deliver better products at a faster pace and with higher levels of customization than traditional approaches.

A broader view

Like other members of the digital workforce, software developers will have to adjust to this new paradigm. One way is to stop thinking like a developer or even an architect and start focusing on the aspects of software that are not so easily automated. One way to do this, says Biz Tech Age’s Harry Miller, is to shift away from perfecting single programs or platforms and look for ways multiple software systems can work together. With the rise of no-code and low-code programs, much of it likely written by AI, human developers will shift away from the practice and process of development to building highly customized solutions for a wide range of challenges.

As AI continues to evolve and acquires the ability to learn on its own and infer from context, we can expect to see broader democratization of software development to the point that virtually anyone can create new programs — no experience necessary. Prem Natarajan, vice president of Amazon’s Alexa program, highlights the way advanced neural networks and other developments are enabling AI to understand plain speech. So instead of having to master a lifetime’s worth of esoteric commands, all anyone will have to do is simply describe what they want and let AI do the rest. Sound too good to be true? We are already seeing this in a rudimentary fashion in the way people are able to reprogram their smart cars, smart homes, and other devices, and this process has only barely begun.

With AI poised to unleash significant disruption across all other areas of the digital landscape, it’s only fair that developers take their lumps. As is true in other areas across the industry, developers must adapt to the changes they have wrought with confidence that things will come out better in the end.