DEVOPSdigest

Will AI replace humans? Will developers be displaced? These are important questions being debated about AI in software development.

"In the early stages of AI development, the use is likely more about productivity increases by accelerating the output of developers," says Chris Du Toit, Head of Developer Relations at Gravitee(link is external). "Inevitably however, this will be utilized to realize cost savings. In a McKinsey study it is estimated that AI could potentially displace between 400 million to 800 million jobs globally by 2030. Developers and IT professionals should assume immunity from this statistic at their own risk."

All in all, AI is not a threat to engineers' jobs, but we're in uncharted territory and will have to wait and see how the job truly changes, adds Camden Swita, Head of AI & ML Innovation at New Relic(link is external).

DEVOPSdigest invited experts across the industry — consultants, analysts and vendors — to comment on how AI can support the software development life cycle (SDLC). In Part 9 of this series, the experts discuss how AI will impact the software development workforce.

Developer Shortage

The 2024 Reveal survey found that the biggest business challenge for the technology industry in 2024 will be working with limited resources (40.9%), followed by recruiting developers with the right skills (34%), cites Casey Ciniello, App Builder, Reveal and Slingshot Senior Product Manager at Infragistics(link is external). "There is a concern that AI might make the developer's role obsolete, however, there is still a great need for skilled developers."

"I have yet to meet an organization that says: we have enough developers to match all the demand for what we could be building," Jeff Hollan, Head of Applications and Developer Platform at Snowflake(link is external), agrees. "There is way more demand for developers than supply, and there is always more we can be building, creating, and enabling with apps and data in businesses for customers."

"I don't see AI replacing developers," Mike Loukides, VP of Emerging Tech Content at O'Reilly Media(link is external), confirms. "When used well, I see it making developers more effective. But I've also never seen a software development shop that didn't have a very long queue of projects they needed to complete, and that wasn't struggling to meet deadlines. AI will take some of the load off of developers — but that will probably mean that the list of projects will get longer and the deadlines tighter."

Copilot, Not Autopilot

Many experts believe that the role of AI is to support, not replace, the developer.

There is a reason Microsoft called its Generative AI tool Copilot and not Autopilot, says Ed Charbeneau, Developer Advocate at Principal, Progress(link is external). It is positioned as an assistant, not as an autonomous replacement.

Jeehong Min, Technical Product Manager at Parasoft(link is external), prefers to think of it as augmentation, not replacement. A developer of any skill level, junior or senior, will benefit from working with an AI tool. It's like having a peer programmer with vast knowledge available on-demand.

Instead of seeing AI as a threat, it should be thought of as complementary support because it allows developers to focus on automating the mundane tasks and instead focus on higher-level design and problem-solving, advises Anand Kulkarni, CEO of Crowdbotics(link is external).

Humans Still Required

"Ah, the allure of AI as a cost-cutting measure! I hear managers say: We can use AI to boost productivity and, more importantly, quality," explains Joanna Schloss, CMO/AI Evangelist at Parasoft(link is external). "While AI, much like any other tool, has its merits,it is by no means a substitute for the people. Indeed, the responsible use of AI necessitates keeping humans in the loop. Developers need the time and focus to engage in thoughtful, high-value coding practices that ensure quality is woven into the very fabric of the code from the outset. Moreover, it is the QA team that must build both automated and exploratory (manual) tests, thereby fostering confidence and minimizing risk throughout the software lifecycle."

Currently, AI can be used to provide a great jumping off point, but it can't fully replace a developer, Udi Weinberg, Director of Product Management, Research and Development, at OpenText(link is external), agrees. For instance, AI can speed up simple tasks, like writing blocks of code for a standard solution, which allows developers to focus on the more creative and strategic aspects of development. At this point, a developer is still required to give it a careful review to ensure it meets the correct needs.

"The conversation around the replacement of workers is one that will never end. Do I think it has the potential to? Sure. Does that mean it should? Absolutely not," asserts Sean Heide, Research Technical Director at Cloud Security Alliance(link is external). "With AI, there will always need to be a human in the middle interaction checking for accuracy, running manual checks, and ensuring that what the system is producing is a fit for the business. The model can only produce what it assumes is the next response in the line of words. Let's remember that. A true 100% reliance on these tools would be just as bad as not having any employees in your company handling daily tasks. Checks and balances is a common statement for a reason. Without it, errors will be made, and businesses will potentially be pushed in a direction they wouldn't need to focus on."

"As LLMs continue to evolve and improve, developers will need to learn to harness this technology in a way that doesn't magnify existing problems," says Todd McNeal, Director of Product Management of SmartBear(link is external). "The most innovative solutions to complex problems in software development will always intersect at the intelligent use of modern technologies and the indispensable human touch."

A balanced approach that integrates AI with human expertise is essential to maintain a robust and secure software development environment, recommends Javed Hasan, CEO and Co-Founder of Lineaje(link is external).

If anything, developers will become even more valuable as their expertise will be required to validate and secure AI-generated code as well as build reusable components, templates, and frameworks to establish governance standards for integrating AI with their existing systems, adds David Brault, Product Marketing Manager at Mendix(link is external).

Creative Thinking

AI is not a major help in design or coming up with good ideas. That's still human territory, says Loukides from O'Reilly Media(link is external). Like Loukides, the experts do not see the developer's role as simply a handler, while AI does most of the work. They see human creativity and problem-solving as an essential component of software development that cannot be substituted.

"AI is at the end of the day only a tool. It lacks imagination and other non-logical human characteristics although it attempts to mimic these," explains Geoff Burke, Community Manager at Object First(link is external). "Developing software is really no different from other human activities; what makes them great and successful is the out of the box surprises and original ideas that come from humans."

Thinking AI will ever replace developers indicates a deep misunderstanding of developers' value. It is not about writing code but knowing what to write under complex constraints, like time, budget, multipart architecture, team resources, and go-to-market goals. It is about creativity and strategy, the coding part is the easiest part, says Mathieu Bellon, Senior Product Manager at GitGuardian(link is external). Developers spend most of their time reading and thinking about what to write, rather than writing the actual code. Maybe a machine will be able to do that in 30 years, but probably at the cost of consuming every GPU available on the planet. One brain, which costs a lot less, will do the same thing and even better.

"AI will replace developers when they replace Dr. Dre, Quinten Tarantino, Homer, and Taylor Swift," Bellon continues. "This is unlikely to happen anytime soon."

"Software development is an outlet for creativity. In the same way the typewriter and personal computer only accelerated the amount of written word in the world, I expect the same pattern for AI and software," adds Hollan from Snowflake.

The integration of AI in software development is not just a technological shift but also a cultural one. It requires a balance between embracing innovation and maintaining human expertise and creativity, says Thomas Fou, VP of Compliance Services at BlueAlly(link is external).

The developer of the future will be more akin to an artist than an engineer, Kyle Carberry, Co-Founder and CTO of Coder(link is external) envisions. AI tools are becoming a staple in software development and over time have progressively lowered the barrier to entry. Soon, anyone can try their hand at coding, much like anyone can attempt to become a painter. However, just like becoming an artist, it requires practice and education.

Changing Workforce

The consensus of experts is that AI is not going to substantially replace developers at first, but it still may change the development workforce as we know it.

Here are some of the ways experts expect AI could impact the workforce:

Loss of Skilled Developers

Where I do potentially see a danger is that AI could replace junior level positions which would then result in there being a lack of experienced developers a few years later. Most people can't simply jump out of college and become a top-level developer. It is in these initial low-level positions where people gain the experience that helps mold them into top-level professionals.

The other problem is that if AI starts to perform many of the basic tasks, new developers will start to rely on this from the very start, and we may end up with a workforce of very shallow workers in the sense of knowledge and understanding of the processes at play. Although, to be fair, I believe that same fear was raised when higher programming languages emerged. The belief that not being able to code at machine level or assembly would leave us with bad programmers did not end up being true. I suspect whenever we introduce new tools there are going to be worries.
Geoff Burke
Community Manager, Object First(link is external)

Smaller Teams

I believe that development teams will likely shrink and gain much more authority. As AI can enhance skill sets and standardize the manual labor of coding, the thought process behind the work will become increasingly crucial. A company that previously required a team of one or two hundred developers may only need a team of twenty or thirty in the future. These smaller teams will be significantly more efficient and will be able to handle all aspects of application development, from ideation to deployment, with the help of AI.
Austin Vance
Co-Founder and CEO, Focused Labs(link is external)

Reduced Reliance on Service Organizations

Companies who adopt AI will be able to do more with the same staff. There should be no immediate impact on current staff, aside from making it easier to eliminate the dead wood. AI will reduce the need to hire additional developers and testers, so it should have the biggest impact on hiring service organizations to perform overflow development and testing.
David Brooks
SVP of Evangelism, Copado(link is external)

Addressing Developer Burnout

AI is already proving to be an incredibly powerful tool for developers, though many feel skeptical about the extent of its capabilities and concerned about the potential it has to disrupt traditional workplace practices, jobs, and processes. I see it as a net positive — by addressing the developer burnout crisis and simplifying workflows, developers can focus on more complex tasks.
Dana Lawson
CTO, Netlify(link is external)

Expanding the Developer Workforce

I predict that AI will help expand the development workforce — democratizing it so a larger set of people will be able to participate in development, while also drastically accelerating the output of the current workforce.
Jeff Hollan
Head of Applications and Developer Platform, Snowflake(link is external)

Go to: Exploring the Power of AI in Software Development - Part 10: Developers