As tech professionals remain hard at work developing artificial intelligence (AI) models, finding new pathways to utilize AI, and discovering where the technology will be most useful, it’s worth looking at how AI will transform the lives of those creating it.

Specifically, how is AI poised to alter the workflow and skillset of tech professionals over the next five to 10 years?

Today, AI exists on shifting sands. New language models pop up daily; GitHub is rife with repos leaning into burgeoning AI technology; Silicon Valley startups quickly see multi-billion-dollar valuations thanks to their unique spins on AI.

Make no mistake: AI is tech’s newest bubble. As with the dot-com boom a generation ago, the bubble will eventually deflate, leaving behind useful technology and a handful of companies smart and tough enough to have weathered the market. We spoke to experts to determine what tech pros should expect from the AI market in 2025 and beyond.

What three changes do you see AI having for engineers and developers before the end of 2025?

The next year promises to be a big one for AI. Though none of the tech pros we spoke with see world-shattering changes in the near term, AI is expected to prove its mettle in 2025.

“By 2025, AI will reshape the role of developers in ways that blend creativity with efficiency,” notes Charlie Clark, founder at Liinks and former Senior Software Engineer at Squarespace.

“First, AI will become the ultimate coding assistant—not just generating snippets but translating high-level concepts into executable code. It will handle the heavy lifting of syntax, allowing engineers to focus on the 'why' rather than the 'how.' This shift will empower developers to create faster without sacrificing quality.

“Second, AI will transform code maintenance. Today, sifting through legacy code is a time-consuming burden, but AI will be able to understand the logic of old systems, refactor them, and even update entire libraries seamlessly.

“Finally, AI-driven automation in CI/CD pipelines will become the norm, optimizing build times and deployment processes based on past performance, making iterative development smoother. It’s not about replacing developers; it’s about letting them do what they do best—solve complex, meaningful problems.”

Sai Chiligireddy, Amazon’s Software Development Manager for Alexa, tells Dice: “AI-powered coding assistants like GitHub Copilot, CodiumAI, and Amazon Q are becoming increasingly sophisticated, with the ability to understand an entire codebase and its surrounding context deeply. These assistants can explain the functionality of specific code segments and provide insights into the overall code flow and architecture. Their capabilities go far beyond syntax checking—they can generate code snippets on demand.

“For example, if an engineer needs to create an Amazon SQS client or set up an AWS EC2 instance, they can prompt the AI assistant, and the code will be generated automatically. These assistants can also detect patterns in the codebase, such as an SQS trigger for an AWS Lambda function, and proactively create the necessary code, permission groups, and associated triggers with minimal manual intervention. While the percentage of auto-generated code making it to production has room for improvement, I expect this to reach a good number, making the coding assistants much more reliable.”

Chiligireddy also sees AI becoming more valuable as a tool for software maintenance, adding: “By the end of 2025, I anticipate that AI assistants will be able to autonomously detect the need for upgrades or security patches, make the necessary changes, and seek approval, streamlining the entire maintenance workflow. Additionally, these AI-powered tools will be able to automatically generate comprehensive documentation and create test cases, further reducing the burden of software maintenance on engineering teams.”

How will engineers and developers use AI in five to 10 years?

Beyond 2025, what do the next five to ten years look like for developers and engineers?

“2024 is already being called the year of AI agents,” Chiligireddy adds. “In the next five to ten years, I anticipate engineers will have access to a suite of specialized AI agents, each focused on a specific domain—one for project planning and risk management, another for design and architecture, a third for coding and optimization, and so on.

“These agents will seamlessly collaborate to deliver end-to-end solutions. Predictive maintenance and proactive issue resolution will also become the norm, as AI constantly monitors the health of software systems and takes corrective actions before problems arise. As a result, the role of engineers will evolve, focusing more on defining the strategic vision, setting the guardrails, and ensuring the AI agents are aligned with the overall business objectives. At the same time, the hands-on coding and maintenance tasks are increasingly delegated to the AI.”

Clark sees a future where AI remains a co-pilot for those writing code—one that might let you know when your code needs to be revamped before you even compile it. “Five years from now, AI will be embedded in every part of the software development process. It will be a technical co-pilot—anticipating bottlenecks before they become problems, suggesting architectural improvements on the fly, and enabling true real-time collaboration. Imagine a future where you explain a feature verbally, and the AI not only drafts the initial code but also builds a prototype environment for testing.

“Fast forward to ten years, and I believe the role of a developer will resemble that of an AI trainer as much as a coder. Developers will craft the core logic but spend just as much time refining AI models to ensure they meet ethical standards and business goals. It’ll be about guiding AI rather than grinding through code, creating a cohesiveness where human insight and machine speed converge.”

What are the pitfalls of using AI as an engineer/developer?

Clark warns that AI can create a “potential erosion of foundational skills,” adding: “If AI automates the basics, there’s a danger that newer engineers won’t build the same deep understanding of core concepts.”

‘Black box solutions’ also concern Clark: “AI models can spit out results that seem perfect, but if developers don’t understand why the AI made those choices, they risk deploying solutions with hidden flaws. It’s like letting a high-speed car drive itself without understanding its braking system—it might work, but when it fails, it fails spectacularly.”

Chiligireddy concurs: “When using AI assistants, it's important to maintain the proper mental model - viewing them as a collaborative co-worker rather than a replacement for one's expertise and decision-making. The engineer or developer should remain the owner of the overall task, leveraging the AI assistant to help with specific sub-tasks and improve productivity.

“Engineers need to be mindful of the limitations and potential biases inherent in large language models (LLMs) used in many AI assistants,” Chiligireddy adds. “These models can sometimes ‘hallucinate’ or produce incorrect outputs, and they may also reflect societal biases present in the data used to train them. It's important to maintain a critical eye, validate the AI's responses, and be aware of these biases, rather than blindly trusting the outputs.”

AI is a tool for developers and engineers. While it can be enticing to let AI do increasingly more for you, overreliance on something you don’t understand is problematic.

Is there an ethical line engineers and developers should consider when using AI?

As we take steps to utilize more AI, we must monitor ethical boundaries.

“The most straightforward ethical boundary is using AI assistants for assessments of individual capabilities, such as in job interviews, academic competitions, or research papers,” Chiligireddy tells Dice. “This would be deceptive and undermine the integrity of the process. A more nuanced consideration is the use of AI in collaborative environments. If an engineer has leveraged an AI tool for tasks like coding, ideation, or system design, there is an ethical obligation to be transparent about it. Failing to disclose the AI involvement would mislead colleagues and erode trust in the collaborative process.”

“In my view, the ethical line in AI development is about transparency and accountability,” Clark adds. “Engineers need to be clear about what AI can do and, more importantly, what it can’t. If an AI model influences a hiring decision, predicts risk in healthcare, or drives any decision that impacts people’s lives, we owe it to users to explain the ‘why’ behind those decisions. But beyond transparency, it’s also about empathy—understanding that the convenience AI offers must never come at the expense of human dignity or fairness.

“We’re not just building products; we’re shaping the experiences of real people. Engineers should approach AI responsibly, ensuring their creations are innovative and just. It’s a delicate balance, but I think that’s where the real potential of AI lies—making technology more human, not less.”