Swarehousing should be artificial intelligence’s easiest victory. Today, companies such as OpenAI, Anthropic, Microsoft and Google have released AI products aimed specifically at coding. And a survey of nearly 5,000 technology professionals released in a report last year by Google’s DevOps Research and Assessment (DORA) team found that 90 percent of respondents said they used AI at work — with more than 80 percent saying the technology had increased their productivity.
“We’re seeing a large majority of people relying on AI to get their jobs done, at least a moderate amount, which is really fascinating,” says Nathen Harvey, who leads the DORA team.
AI can generate code for everything from web and mobile apps to data management tools. It often automates some of the tedious elements of the job, such as building test infrastructure and updating software to run on new devices and systems. In some cases, even inexperienced developers can create working prototypes simply by describing their intentions to AI systems in a process often called “vibe coding,” a term coined by OpenAI founder and researcher Andrej Karpathy. But writing code is only part of the job; Developers still have to verify that it does what it’s supposed to and fix it if it fails.
On supporting science journalism
If you like this article, please consider supporting our award-winning journalism by subscribes. By purchasing a subscription, you help secure the future of impactful stories about the discoveries and ideas that shape our world today.
Another finding from the DORA report was that while the effectiveness of individual code appeared to increase with the use of AI, so did “software delivery instability” – a measure of how often code had to be rolled back or patched after release to resolve unexpected problems.
“When you use more AI, you’re more likely to roll back changes you’ve pushed into production,” says Harvey. “And this is obviously something you want to avoid.”
Although it is becoming increasingly adept at writing code, AI does not eliminate the need for human software development. Developers often still need to create custom code—or at least adjust the output of an AI tool—to handle unusual cases or specific business needs that may not be reflected in AI training data. They must also still verify that machine-generated programs behave exactly as intended and meet company standards.
AI tools do not automatically shorten the working day. In some workplaces, studies show, AI has increased the pressure to move faster than ever.
If employers don’t manage its effects, AI may even exacerbate stress and burnout among software engineers. In a report published in Harvard Business Review in February, researchers at the University of California, Berkeley’s Haas School of Business found that employees at one US technology company took on more tasks, worked at a faster pace and worked longer hours after adopting AI. Even without the company mandating use of the technology, employees began calling on AI during lunch, breaks and meetings, and some found earlier downtime less refreshing. There is a risk that initial excitement and productivity increases may give way to fatigue, lower quality and greater employee turnover, the researchers warned.
This pressure does not occur in a vacuum. After years of industry-wide layoffs and corporate mandates for efficiency, AI is often deployed alongside the expectation that those left behind will do more with less.
Additionally, a report assessing more than 500 developers released late last year by Multitudes, a New Zealand-based company that helps companies track and optimize software development practices, found indications that AI could expand worker productivity, but also work hours. On average, engineers merged 27.2 percent more “pull requests” – packages of code that were approved for insertion into existing software projects. But they also experienced a 19.6 percent increase in “out-of-hours commitments” — submissions of coding work outside of their regular schedules. It could be a sign of trouble to come.
“If it increases outside of work hours, it’s not good for the person,” says Multitude founder and CEO Lauren Peate. “It can lead to burnout.”
The Multitudes report doesn’t definitively prove that AI directly caused the measured changes, but Peate says interviews suggest that the observed changes in hours among engineers are likely a sign that companies expect greater productivity from employees in the AI era.
“People felt extra pressure to get more work done, and it looks like that helped them put in more hours,” she says.
Although some research has suggested that less-experienced developers may be among those who benefit most from AI’s assistance, and vibe coding may allow people with minimal programming background to build programs that run, a recent assessment by Anthropic suggests that over-reliance on AI may affect the development of coding skills.
In a report released in January, Anthropic researchers found that software engineers working with a new software library saw a small, statistically insignificant increase in speed when solving a task with the help of AI compared to a control group working without AI assistance. When the coders were asked about the software library after the task, the AI-assisted group scored 17 percent lower than the AI-free group. Those who asked questions about AI rather than just relying on it to generate code generally performed better, but the researchers raised concerns that using AI to complete tasks as quickly as possible under work pressure could be detrimental to engineers’ professional development.
In addition, they noted, the biggest gap in quiz performance was on questions related to debugging code — the process of finding and fixing the bugs that make code not work. In other words, junior developers who rely too much on AI may have a harder time not only writing code on their own, but also understanding and finalizing the code they generated in the first place. In a statement to Scientific American, Anthropic researcher Judy Hanwen Shen said the goal “shouldn’t be to use AI to avoid cognitive effort—it should be to use AI to augment it.”
Already, the UC Berkeley researchers noted, engineers may find themselves helping coworkers who have created incomplete software solutions through vibe coding. And some open source projects have reported an increase in low-quality, AI-driven submissions that eat away at core developers’ time.
It comes after a 2025 Harvard Business School working paper indicated that AI could lead to open source developers shifting their time from handling project management tasks, such as reviewing code contributions and maintaining lists of issues that contributors can fix, to generating code themselves.
“You can do it yourself now, so there’s not a lot of need to interact a lot with others,” says Manuel Hoffmann, a co-author of the paper and an assistant professor of information systems at the University of California, Irvine’s Paul Merage School of Business. “And that’s not necessarily a bad thing.”
Still, such use of AI can limit another channel for less experienced programmers to hone their skills, develop professional networks, and expand their resumes.
And as AI redefines what productivity means, workplace structures that prevent burnout, keep workloads manageable, and provide opportunities for advancement and training may be more important than ever.
“When you have big things happening, and you add some AI to the mix, they’re probably going to get better,” says Harvey. “And when you have painful things happening, (and) you add some AI to the mix, (you’re) likely to feel that pain a little more acutely.”
