da Agency

Docker State of App Dev: AI

AI is changing software development — but not how you think

The hype is real, but so are the challenges. Here’s what developers, teams, and tech leaders need to know about AI’s uneven, evolving role in software.

Rumors of AI’s pervasiveness in software development have been greatly exaggerated. A look under the hood shows adoption is far from uniform. While some dev teams are embedding AI into daily workflows, others are still kicking the tires or sitting it out entirely. Real-world usage reveals a nuanced picture shaped by industry, role, and data readiness.

Here are six key insights into AI tools and development from Docker’s second annual State of Application Development Survey, based on responses from over 4,500 industry professionals.

1. How are people using AI?

Right off the bat, we saw a split between two classes of respondents: 

Those who use AI tools like ChatGPT and GitHub Copilot for everyday work-related tasks such as writing, documentation, and research 

Those who build applications with AI/ML functionality

2. IT leads the way in AI tool usage and app development 

Only about 1 in 4 respondents (22%) report using AI tools for work. But there’s a huge spread across industries — from 1% to 84%. Among the top AI users are IT/SaaS folks (76%). And because we surveyed over three times more users this year than for last year’s report, the snapshot covers a broader spectrum of industries beyond just those focused on IT.

Underscoring tech’s embrace of AI: 34% of IT/SaaS respondents say they develop AI/ML apps, compared to just 8% outside that bubble.

And strategy reflects this gulf. Only 16% of companies outside IT report having a real AI strategy. Within tech, the number soars to 73%. Translation: AI is gaining traction, but it’s concentrated in certain industries — at least for now.

3. AI tools are overhyped — and incredibly useful

Here’s the paradox: 64% of users say AI tools make work easier, yet almost as many (59%) think AI tools are overhyped. The hype may be loud, but utility is speaking louder, especially for those who’ve stuck with it. In fact, 65% of current users say they’re using AI more than they did a year ago, and that same percentage use it every day.

This tracks roughly with findings in our 2024 report, in which 61% of respondents agreed AI made their job easier, even as 45% reported feeling AI was overhyped. And 65% agreed that AI was a positive option.

4. AI tool usage is up — and ChatGPT leads the pack

No surprises here. The most-used AI-powered tools are the same as in our 2024 survey — ChatGPT (especially among full-stack developers), GitHub Copilot, and Google Gemini. 

But usage this year far outstrips what users reported last year, with 80% selecting ChatGPT (versus 46% in our 2024 report), 53% Copilot (versus 30%), and 23% Gemini (versus 19%).

5. Developers don’t use AI the same way

The top overall use case is coding. Beyond that, it depends.

Seasoned devs turn to AI to write documentation and tests but use it sparingly. 

DevOps engineers use it for CLI help and writing docs.

Software devs tap AI to write tests and do research.

And not all devs lean on AI equally. Seasoned devs are the least reliant, most often rating themselves as not at all dependent (0/10), while DevOps engineers rate their dependence at 7/10. Software devs are somewhere in the middle, usually landing at a 5/10 on the dependence scale. For comparison, the overall average dependence on AI in our 2024 survey was about 4 out of 10 (all users).

Looking ahead, it will be interesting to see how dependence on AI shifts and becomes further integrated by role. 

6. Data is the bottleneck no one talks about

The use of AI/ML in app development is a new and rapidly growing phenomenon that, not surprisingly, brings new pain points. For teams building AI/ML apps, one headache stands out: data prep. A full 24% of AI builders say they’re not confident in how to identify or prepare the right datasets.

Even with the right intent and tools, teams hit friction where it hurts productivity most — upfront.

Bottom line:We’re in the early stages of the next tech revolution — complex, fast-evolving, and rife of challenges. Developers are meeting it head-on, quickly ramping up on new tools and architectures, and driving innovation at every layer of the stack. And Docker is right there with them, empowering innovation every step of the way.

Quelle: https://blog.docker.com/feed/