DevOps
Most software engineers now use AI in most of their code and are concerned about existential threats.
According to the State of Web Development AI survey, nearly half of web developers are concerned that AI will replace their jobs, with one saying, “AI will have a devastating impact on our field.”
This survey of 7,258 developers is the second conducted on this topic by Devographics, which has other surveys such as State of JavaScript and State of CSS.
There has been a significant shift since the first survey in early 2025, when the majority of respondents used AI to create less than 25 percent of their code, to now 63 percent of developers use AI to generate more than half of their code. More than a quarter (27%) of respondents use AI in more than 90% of their code.
Code generation is the most common use case for AI, followed by code review, investigation, and debugging.
The researchers drew respondents from those who had responded to previous surveys as well as those contacted through social media, and said this theme may have “biased respondents toward developers interested in AI.”
When it comes to job security, the general consensus is that a developer’s skills will continue to matter in the world of AI, but bosses can persuade them otherwise and fire them.
“AI companies can convince employers that AI can take my job, even if they can’t,” said one person. Another commented, “I’ve already had to look for something new because my job as a designer and front-end developer was canceled because of AI.”
The number of young employees hired is decreasing, and there are concerns about a decline in skills. “Companies would rather spend money on AI than train their employees,” one person commented.
The most used model provider is ChatGPT (88.4 percent), slightly ahead of Anthropic’s Claude (82.1 percent). However, when it comes to paid subscriptions, Claude is the winner (69 percent), followed by ChatGPT (49 percent) and Google Gemini (32 percent).
Although usage is on the rise, respondents are by no means AI enthusiasts. The use of AI to generate images has fallen from 38 percent to 37 percent since last year, with some respondents objecting to it on ethical grounds.
“I don’t use image generators as a rule,” one said, while another claimed, “AI image generators are built entirely on stolen images.”
The general section on AI risks revealed a number of concerns. At the top of the list was separation, but not far behind were military uses of AI, environmental impact, and AI slop takeover. Security issues and rising costs were also areas of concern. The survey limited respondents to their top three choices. Many comments indicated that they would have liked to have chosen more.
From a technical perspective, the biggest problems cited were illusions and inaccuracies (64%). Poor code quality (53%), lack of context (38%).
This is a strangely complex situation, with respondents expressing strong concerns about the overall impact of AI, while at the same time becoming increasingly reliant on it. 74% agree that AI tools are essential to their workflow, and 64% feel that AI has increased their productivity. 88% believe the quality of AI tools has improved significantly over the years. ®
