AI at Work: How Software Engineers Are Using Artificial Intelligence on the Job

Since the dramatic rise of modern AI usage beginning just a handful of years ago, questions about its reliability, effectiveness, and potential to take over our jobs have abounded. But today, in the world of software development, employers are increasingly recognizing artificial intelligence’s value.

According to our new State of Remote Engineering report, the vast majority of engineers—a staggering 80%—say they are encouraged to use AI on the job. 

How are engineers and businesses embracing AI rather than fearing it? What skills and tools are proving most useful in this uncharted territory? What are the perceived benefits and risks of relying on AI in software development? Let’s take a look at AI at work.

AI in Software Development: Use Cases

We know AI is affecting nearly every field, none perhaps more so than software development, the very field that created the technology as it exists today. According to a Wired survey, three in four coders have tried AI, and most use it on a weekly basis. 

So, just how are tech professionals using it?

Our report found that the most prominent use case is an AI coding assistant. Nearly two-thirds (64%) of the engineers we surveyed leverage these tools. So, we can glean that greater productivity is in high demand. AI coding assistant or copilot usage can also speak to the pressure to accelerate the pace of development and reduce time to market. Automation of repetitive tasks—arguably similar to AI coding assistance / copiloting—is less commonly cited in our survey, although 39% of engineers say they use AI in this way.

Our survey also reveals that many engineers leverage AI tools for refinement, with 50% saying they use them for code review and 38% using them for debugging. Use cases are fairly widespread. For example, many engineers also use AI tools for code review (50%) and documentation (50%), as well as research and development (44%), among others.

When considering the available AI tools, more than a third (35%) prefer ChatGPT, with no other tool coming close to being as popular. Other options in our survey included Copilot (16%), Gemini (14%), and Claude (8%), along with several additional choices. This speaks to GPT’s overwhelming pervasiveness across industries, not just software development. And, of course, our survey took place before the release of GPT-5 in August 2025, which promises to gain even more traction for the LLM.

Is AI Helpful or Hurtful?

While there is no doubt that engineers are using AI on the job, their feelings about the technology are mixed.

Wired found that newer professionals are more likely to support AI usage. Three-quarters of those who have been coding for under a year say they are “AI optimists.” Meanwhile, mid-career coders are less AI-friendly: Nearly half of those surveyed call themselves “AI pessimists.” 

Our survey is consistent with these findings, showcasing conflicting views on AI. For example, while 37% of respondents are extremely or very confident in the quality and efficacy of AI-generated code, 47% are only “somewhat” confident. Seventeen percent are not confident.

Want all the latest insights from 1,000s of global engineers? Get the State of Remote Engineering

Stack Overflow, meanwhile, found that while more developers (76%) are using AI coding tools or plan to, confidence in these tools is down—72% compared with 77% the previous year. Many (31%) simply don’t trust the accuracy of AI output; 42% do trust the accuracy, and 27% are noncommittal.

Still, the Stack Overflow survey does show many benefits, saying AI tools increase productivity (81%), speed up learning (62%), and improve efficiency (58%). Notably, just 30% claim that they improve accuracy in coding. 

So, while opinions on AI tools are mixed, it is clear that these tools do have many benefits for software engineers—along with several drawbacks, of course. 

There are some real risks and limitations associated with AI-generated code and AI development tools in general. These range from security vulnerabilities and compliance issues to a lack of governance and oversight.

Given the mixed confidence in AI for software development purposes and the real risks of using artificial intelligence for coding, why should engineers rely on it at all? Well, there are still the numerous benefits we’ve underscored.

However, it’s clear that AI tools need to be paired with rigorous vetting processes on the part of engineers, who must recognize the limitations of these tools and not rely on them blindly. 

Moreover, AI can prove valuable as a teaching tool. The research study “AI in software programming: understanding emotional responses to GitHub Copilot” finds that AI tools like GitHub can support educational programs, helping newbie programmers gain skills and use technological tools knowledgeably and successfully. 

Conclusion: What Is the Future of AI in Software Development?

Much has changed in just the past year when it comes to AI at work. 

McKinsey estimates generative AI could add $2.6–4.4 trillion to the global economy. And it won’t just affect software engineers’ jobs—it will also automate repetitive tasks for project managers, product managers, UX/UI designers, QA engineers, and other tech professionals. However, as Thomas Dohmke, CEO of GitHub, notes, “Knowing how to figure out whether the content provided by the AI is actually the right answer is going to be crucial.”

With many employers buying into the innovative power of the technology, focusing on learning how to use these ever-changing tools productively and efficiently remains pivotal. Software engineers who have not yet embraced AI could very well fall behind. 

While there is some fear surrounding artificial intelligence tools, understanding their limitations and advantages proves indispensable in software development best practices. 

It’s clear that nearly all engineers understand that AI will transform the way they work. Now, it’s up to engineers to use it as a partner in their efforts—and use it responsibly.